The Progress of Medical Image Semantic Segmentation Methods for Application in COVID-19 Detection

Valizadeh, Amin; Shariatee, Morteza

doi:https://doi.org/10.1155/2021/7265644

Computational Intelligence and Neuroscience

On this page

Abstract Introduction Discussion Limitations Conclusion Data Availability Conflicts of Interest Acknowledgments References Copyright Related Articles

Special Issue

Swarm Intelligence and Neural Network Schemes for Biomedical Data Evaluation

View this Special Issue

Review Article | Open Access

Volume 2021 | Article ID 7265644 | https://doi.org/10.1155/2021/7265644

The Progress of Medical Image Semantic Segmentation Methods for Application in COVID-19 Detection

Amin Valizadeh¹and Morteza Shariatee²

Academic Editor: Suresh Manic

Received14 Apr 2021

Accepted18 Oct 2021

Published22 Nov 2021

Abstract

Image medical semantic segmentation has been employed in various areas, including medical imaging, computer vision, and intelligent transportation. In this study, the method of semantic segmenting images is split into two sections: the method of the deep neural network and previous traditional method. The traditional method and the published dataset for segmentation are reviewed in the first step. The presented aspects, including all-convolution network, sampling methods, FCN connector with CRF methods, extended convolutional neural network methods, improvements in network structure, pyramid methods, multistage and multifeature methods, supervised methods, semiregulatory methods, and nonregulatory methods, are then thoroughly explored in current methods based on the deep neural network. Finally, a general conclusion on the use of developed advances based on deep neural network concepts in semantic segmentation is presented.

1. Introduction

Semantic segmentation of medical images is also known as pixel-level classification. The task is to cluster the parts of the image side by side, which belong to a class of similar objects [1]. The other two key functions of the image are to classify the image’s surface and define it. Image classification ensures that each image is exchanged as an equal group of images of similar groups, and monitoring also refers to the object’s location and recognition. For predicting pixel level, image segmentation can be used as it categorizes each pixel. Furthermore, there is a task that identifies and separates joints called sample segmentation [2, 3]. Medical image semantic segmentation has a variety of applications, such as road sign detection [4], colon crypt segmentation [5], land-use classification, and land surface classification [6]. It is also widely used in medicine, such as brain and tumor detection [7] and discovering and tracking medical devices in surgery [8]. Numerous applications of segmentation in medicine are listed in some studies [8]. Scene resolution is of great significance in advanced driver assistance systems (ADAS) or car driving areas and depends extensively on semantic image segmentation ([9–11]). Research has developed a deep-learning (DL-) based system for assessing disease. This system automatically scans the location of the disease and measures the shape, size, and percentage of the disease on the CT image of people who have COVID-19 disease. In this study, a strategy (HITL) was proposed for the repeated production of training samples. This method is for radiologists to evaluate the results of DL segmentation, make changes, and frequently add more tutorials to update the model. As a result, they speed up the algorithm’s development cycle [12]. As recent findings show, before choosing to use chest CT, a significant number of imaging studies need to be checked for patient diagnosis or patient screening. Artificial intelligence technology, particularly DL analysis tools, could potentially be created to support radiologists in triage, quantification, and data analysis. Artificial intelligence solutions can analyze several cases to determine if a chest CT scan shows lung abnormalities. If the software significantly increases the risk of developing the disease, the case will be reviewed by a radiologist or a physician for further treatment/quarantine. Such systems or their modifications, after validation and testing, can be a key factor in the diagnosis and control of patients with the virus [13]. The pandemic of COVID-19 appears to have negative impacts on world health and well-being. An important method in combating COVID-19 is efficient screening in infected patients, and one of the most important screening methods is radiological test utilizing chest radiography. Preliminary studies have shown that patients infected with COVID-19 have problems with chest radiographs. COVID-Net is implementing a deep concealer neural network architecture to diagnose COVID-19 patients implying chest X-ray (CXR) images motivated by the academic community’s open-source endeavors. This information is accessible to the public [14]. In the context of e-healthcare, Zhang et al. showed a privacy-preserving optimization of the clinical pathway query method (PPO-CPQ) [15]. Ala et al. have used a metaheuristic algorithm and optimized an appointment scheduling issue for healthcare systems depending on the quality of fairness service [16]. Also, Xu et al., to simulate pathogenesis diagnosis, proposed a computer technique called network differentiation [17]. Segmentation accuracy has greatly improved since the reemergence of the deep neural network. In general, traditional methods are called the methods that came before the deep neural network. The following parts of this convention are followed in this study and standard segmentation techniques are briefly analyzed in this article, and, most significantly, this development builds on the recent progress of adopting and organizing a deep neural network from different aspects. Furthermore, the image segmentation measurement and assessment databases are checked. The rest of this study is organized as follows: In the dataset and assessment criteria, Section 2 explores the semantic segmentation of the image. A description of traditional methods is given in Section 3. Section 4 outlines recent developments in detail. Finally, a description of the work performed is given with conclusions in Section 5.

2. Datasets and Evaluation Metrics

2.1. Datasets

Many general datasets are currently connected to image segmentation, such as PASCAL VOC, MS COC, ADE20K, and, in the field of the autonomous driving area, Cityscapes [11] and KITTI ([9, 10]). The challenge of visual object classes, or VOCs, consists of two components [15]: (1) image and annotation datasets that are available to the public and (2) annual workshops and competitions that are held online on some websites and sometimes in person. The main challenges have been dealt with since 2005. By 2012, the challenge included 20 classes. Educational and validation data contained 11530 images containing 27450 annotated objects with areas of interest and 6929 segmented images. Also, in image segmentation, datasets have been extensively utilized. The Microsoft COCO dataset [2] contains images of 91 objects, where a 4-year-old person can quickly identify with 2.5 million labeled samples in 328,000 images. Authors also introduced the dataset with a detailed statistical analysis compared to PASCAL data [15], ImageNet data [18], and SUN data [19]. An analysis of 50 chest X-ray images of 25 positive COVID-19 cases was confirmed because of the lack of an available COVID-19 dataset. Seven distinct architectures from neural network models are used in COVIDX-Net. Each deep neural network model can analyze the number of X-ray images to identify the patient’s state as negative or positive COVID-19 [20]. The authors in [20] collected images from 5 different sources to test this idea and generate a dataset of 170 X-ray images and 361 CT images of COVID-19. Two explanations exist for the use of photographs from these sources. First, to help radiologists diagnose COVID-19 worldwide, it is important to design an advanced tool. Second, for the scientific community and the general public, photographs of these sources are openly available. Also, the images used in it will be publicly available in a GitHub repository [21]. With 150 classes of objects and materials, ADE20K data [17] is another scene analysis criterion. ADE20K data contains the object segmentation mask and component segmentation mask, unlike other datasets. There are also several pictures of parts of the head (like the mouth, eyes, and nose). In the training suite, there are precisely 20210 images, and there are 2,000 images in the validation suite and 3,000 images in the experimental suite [17]. Some of these images are depicted in Figure 1.

The Cityscapes dataset is a criterion that focuses on understanding the meaning of urban street scenes [11]. The collection contains 30 groups obtained from 50 towns in 5,000 fine annotated pictures. The selection period, which includes spring, summer, and autumn, is also several months. Figure 2 displays one of the images of this data in the annotations.

The KITTI dataset [9], another autonomous driving dataset recorded by driving on highways and in rural areas around Karlsruhe, is another example of semantic image data. On average, a maximum of 15 cars and 30 pedestrians can be seen in each image. Zhou et al. proposed a model for evaluating the clarity of screen content and natural scene images while blind [10]. Lv et al. proposed a deep-learning-based fine-grained visual computation [11]. Liu et al. have investigated the Style and Characters Inpainting Based on CGAN [12]. Road detection, stereo reconstruction, light current visual measurement, 3D object detection, and 3D tracking are the principal functions of this dataset (http://www.cvlibs.net/datasets/kitti/). One use of image segmentation is in automated vehicles. The system uses augmented reality to describe the amount of automation and its dependability to increase the system's confidence and reliability. In addition to the above databases, there are many others such as SUN, the Visual Database of Shadow Detection or Texture Segmentation (https://zenodo.org/record/59019#.WWHm3oSGNeM), Berkeley segmentation dataset [22], and LabelMe dataset [23] whose complete information can be found (http://homepages.inf.ed.ac.uk/rbf/CVonline/Imagedbase.htm). There are various imaging models in the field of medical data, most of which have been applications of DL methods on MRI, mammography, or CT scan imaging data. However, different areas of the body have this data or even other imaging samples. However, the main focus of this research is on three datasets, which are from the brain and chest areas. One of the most important brain MRI datasets working on diagnosing tumors, Alzheimer’s, and MS is BraTS dataset (https://ieee-dataport.org/competitions/brats-miccai-brain-tumor-dataset). From the download path of this dataset, there will be four folders: T1, T2, Flair, and T1Ce, respectively. Each section has 155 sections, MRI image sections are weak, and there are 155 sections for a dataset and 210 sections for a high-grade Glioma dataset. At the same time, 75 sections are in another type of Glioma. Hence, the number is 285 cases. This dataset has versions between 2012 and 2019, and an example of the images of this dataset is in form (3). There is also another original dataset called TCGA-GBM, which has higher quality 3D images than BraTS (https://portal.gdc.cancer.gov/projects/TCGA-GBM) (see Figure 3).

There are also valid datasets in the field of mammography images used to diagnose breast tumors, one of the most important of which is a dataset called MIAS (https://www.mammoimage.org/databases).

Also, on the same website, other datasets called DDSM, AMDI, and IRMA are used to diagnose cancerous tumors in different shapes and sizes as benign and malignant. An example of these images is shown in Figure 4.

2.2. Evaluation Metrics

For image segmentation and scene analysis, standard performance assessment metrics include pixel resolution , middle resolution , region intersection upon the union , and connection area sharing frequency weight . It is presumed that describes the number of class i pixels that are supposed to belong to class j, where there are various groups and it is assumed that represents the number of pixels in class i. All of these relationships are written in the four following formulas [26]:

There are also other evaluation metrics for the segmentation of medical images, which are popularly used in scientific societies. These include accuracy, sensitivity, specificity, recall, ROC curve, and Area under Curve (AUC) rate.

3. Traditional Methods

Before presenting deep neural networks, features and classification methods are applied to the most important topics [19]. A feature is a piece of data that is applied to solving computational tasks in machine vision and image processing. That is the same context of machine learning precision and the identification of patterns [27]. For semantic segmentation in images, a number of features are used, such as pixel color, histogram of oriented gradients (HOG) ([28]), scale-invariant feature transform (SIFT) [29], local binary pattern (LBP) [30], SURF method [31], Harris Corner Detection [32], method of Shi-Tomasi [33], subpixel corner method [34], SUSAN edge method [35], Features from Accelerated Segment Test (FAST) [36], FAST-ER method [37], AGAST method [38], AGAST multiscale detection method [39], the bag-of-visual-words (BOV) [40], the Poselets method [41], the Textons method [42], and many other methods. Approaches to image semantic segmentation do not include supervised or unsupervised cases [43]. In particular, thresholding, which is commonly used in gray surface images, is a simple method. In the medical industry, optimization, classification, and diagnosis are very common, using imaging equipment ([44–46]). In general, in this regard, thresholding methods are very efficient. K-means clustering means an unsupervised clustering process. The K-means algorithm specifies that the number of clusters must be defined in advance. The points of k are initially randomly positioned in the property space. Additionally, each datum is allocated to the closest points. The gravity points are then successively transferred to the middle of the cluster. This process proceeds until the stop criterion is met [47]. The problem of segmentation can be used as an energy model, resulting from a compression method [48]. Intuitively, edges are an important part of segmentation, and there is also much research on edge recognition ([49–53]). Also, edge-based approaches and regional growth methods are other branches. Support vector machines (SVMs) are binary classifiers that are well studied and are employed in many jobs. Inseparable linear problems can be solved with the slack variables, too [54, 55]. The core approach was also used for integral tasks by the mapping of dimensional broader features. A Markov random field (MRF) is a set of randomized variables with an indirect diagram of a Markov attribute. Also, the Markov stochastic grid is a directionless graphical model (http://host.robots.ox.ac.uk/pascal/VOC/voc2010/results/index.html; [56]).

In general, a case study has been made between the methods of segmentation in images in the field of semantic segmentation, which is shown in Table 1. Classic studies on the detection of cancerous tumors from MRI images have also been presented. In [57], the method of Brownian motion of water molecules to produce contrast has been done. Also, in [58] an improved edge detection method for segmentation is presented. The watershed method and the hierarchical clustering algorithm have also been studied in [59]. Also, in [60], anisotropic diffusion based on segmentation and pattern based on group classification based on support vector machine and segmentation with FCM has been done. The application of a genetic algorithm and discrete wavelet transform thresholding method is presented in [59]. Qiao et al. have presented a local wavelet acoustic pattern and an MLP optimized by a modified Whale Optimization algorithm for classification of underwater objects [59].

Also, [65] presents a combined approach called ant colony optimization (ACO) algorithm and genetic algorithm. In [66], the chaotic firefly algorithm based on the FCM algorithm has been performed. The application of the optimal Particle Swarm Optimization (PSO) in [67] has also been studied, and in [68] a bat optimization algorithm for segmenting MRI images for different purposes is presented. In general, a case comparison between the existing methods in the segmentation field in MRI images has been done, which is shown in Table 2.

In the field of classical methods used to segment mammographic images, we can refer to the research [69] that has used the segmentation method of regional growth with a cellular neural network with a specific threshold. The use of a Back-Propagation (BP) neural network has also been considered in [70]. Applying the new Naïve Bayesian method in [71] has also been considered in this field. In [72], the regression-based evolutionary methods are used to diagnose breast cancer tumors to estimate and predict the remaining life based on the size of the tumor. In [73], the classification or diagnosis of breast cancer in mammographic images combined with wavelet analysis and genetic algorithm is presented. Xu et al. provided a method for identifying, classifying, and predicting nucleic acid-binding proteins [74].

Also, [75] presents a semisupervised adaptive algorithm named GrowCut for the segmentation of tumors of interest areas or ROI of mammographic images based on the amendment of the law of automatic evolution. In general, a case comparison has been made between the methods available for segmentation in mammographic images, which is shown in Table 3.

4. Recent Deep Neural Network Methods in Segmentation

4.1. Artificial Neural Network (ANN)

Biological neurons are inspired by the artificial neural network (ANN). An artificial neuron is an essential element of an artificial neural network. Each artificial neuron has only inputs that weigh. Neurons issue a scale following a transfer function or activation function. An instance of a neural model is shown in Figure 5.

Based on artificial neurons, the accumulation of different neurons is automatically Autoencoder [81], Restricted Boltzmann Machine (RBM) [82], Recurrent Neural Network (RNN) or convolutional neural network (CNN) [83], Long Short-Term Memory (LSTM) [84], and other models. The basic architecture is shown in Figure 6.

A shared weight architecture, influenced by biological mechanisms, is used by the convolutional neural network (CNN) [83]. The connection pattern between neurons mimics the development of the visual cortex of the animal. Acceptance is another essential term, indicating that individual cortical neurons can respond to stimuli only in a small region of the visual field. They also have immutable or complex spatial properties dependent on the architecture with shared weight and spatial characteristics. Due to this excellent structure, the convolutional neural network has gained significance which caused image classification, segmentation, and detection. In the following section, recent developments using convolutional neural networks in the semantic segmentation of the image will be presented.

4.2. Fully Convolutional Network (FCN)

The article in [86] represents the first study in the image segmentation field to present ANNFCN. Replacing the utterly connected layer with the fully convolutional layer is the fundamental concept of this method. Using the interpolation layer, the network recognizes that the output size is just like the input required for segmentation. Most significantly, by successful inference and learning, the network is educated, takes on the required size, and produces the correct size output. The FCN was introduced in VGG-Net and has reached a substantial role in the segmentation of PASCAL VOC (20 percent relative increase to 62.2 percent of the average IU in 2012). However, the assumption takes less than one-fifth of a second for a regular image. The main FCN architecture is shown in Figure 7.

4.3. Interpolation against Parsing in Medical Image Semantic Segmentation

The parsing layer is also approved in the semantic segmentation of images and the FCN architecture. Degradation and sampling layers in the pooling layer that define pixel type labels and predict segmentation masks are the degradation network used in [88]. Unlike FCN in [88], this grid is used for proposed thing designs to get the synthetic parts as an example for the final semantic segmentation. The sampling method step adopts two-line interpolation, which can be found in [86]. The sampling stage of the samples has commonly approved two-line interpolation due to the computational efficiency and good retrieval of the original image. The decomposition operation is an inverse calculation of the convolution function that can also retrieve the input size. It can then be utilized to segment the function mapping size to the original input size to retrieve it. The architecture used in [88] is seen in Figure 8 . Some researchers still use the decomposition layer in multiple versions to introduce semantic segmentation, which can be found in [74, 89, 90].

CT images are also used to obtain information about COVID-19 patients. The CT image shows the condition of the patients' lungs and shows how much the disease has affected the lungs.

4.4. Connect FCN with CRF and Other Traditional Methods

Responses in the last layer of deep convolutional neural networks (DCNNs) are not sufficiently localized to effectively segment an entity, according to DeepLab research [92]. Mixing an ultimately linked random field or CRF in the DCNN end layer solves this poor localization property. Authors’ method in the test determined in the semantic image segmentation work of PASCAL VOC-2012 reaches 71.6% IOU accuracy. After this, they implement another segmentation architecture by matching Domain Transform (DT) with DCNN [92]. Since the dense CRF inference is costly in terms of computation, Domain Transform (DT) relates to a modern method of maintaining edge filtering. A reference edge mapping governs the smoothing rate. The Domain Transform (DT) is several times more rapid than the dense CRF assumed. Finally, studies compare the effects of semantic segmentation and reliably document the boundaries of the object. Researchers also use superpixels to segment images in the domain of the semantic segmentation [93]. Reference [94] deals with semantic segmentation by merging rich details, including mixing label fields and high-order relationships, in the Markov random field (MRF).

4.5. Dilated Convolution

The majority of semantic segmentation methods focus on the compatibility of convolutional neural networks (CNNs), initially designed to classify images. Dense prediction, however, is structurally distinct from classification, as are semantic image segmentation tasks. An instance of an open structure of convolution can be seen in Figure 9. Reference [92] has previously used this technique in its work, which is named Atheros convolution or convolution hole [92] or open convolution [96]. Convolution was originally developed to efficiently calculate wavelet transform in an “algorithm à trous” scheme [97]. Reference [96] systematically presented a module using discrete complexities for collecting multiscale textual information. This architecture is based on open convolution, which supports the expansion of the exponential receiver field without losing sharpness or coverage, since the available convolution contains networked (or segmented) network artifacts in the input data.

(a)

(b)

(c)

Reference [98] created an approach called dilated residual networks (DRN) to eliminate these artifacts and further improve network efficiency.

4.6. Progress in the Main Pillar of the Network

The network’s main column refers to the network’s main structure. As it turns out from image classification tasks, the key source of image medical semantic segmentation is derived. The FCN [86] approved the VGG-16 net structure [99], which performed exceptionally well in ILSVRC14. Authors also considered the architecture of AlexNet [100], which won ILSVRC12, as well as GoogLeNet [101], and performed well in ILSVRC14. The VGG network has been validated in many previous studies [92, 94]. Following the release of ResNet or the deep residual network [102], DeepLab took its place in the ILSVRC 2015 classification work and implemented it, and semantic segmentation has made new progress. To reach a sufficient configuration, [103] evaluates the various changes of a fully complex residual network, including feature mapping resolution, number of layers, and field of view size. Also, [104] examines the remaining deep networks and explains some of the experimentally observed habits. As a result, authors get a shallow network architecture that in the ImageNet classification dataset is dramatically better than much deeper ones. Recently, ResNeXt [105] was introduced as the next generation of ResNet. This network is the basis for entering the ILSVRC 2016 classification work, in which it has won second place. GoogleNet also acquires extensions such as Inception-v2, Inception-v3 [106], Inception-v4, and Inception-ResNet [101], which has already been approved in the article [107].

5. Pyramid Methods in Semantic Segmentation

Aside from adopting networks with strong core columns, researchers are also trying to combine a pyramid strategy with CNN. A clear example of that is a pyramidal structure.

5.1. Image Pyramid

An image pyramid [108] is a series of sequentially segmented images before any of the desired stop criteria has been met. There are two different types of image pyramids: the Gaussian pyramid used for image sampling and the Laplacian pyramid used to recreate a scattered image from the lower image (with lower resolution). Three levels of the image pyramid can be seen in Figure 10.

In the field of image semantic segmentation, [110] establishes a network that can efficiently boost output with traditional multiscale image input and sliding pyramid mixing. This architecture captures the sense of the background patch. Similarly, by feeding input images of various sizes into a deep sharing network, DeepLab implements an image pyramid structure [111] that extracts multiscale functionality. The resulting features are combined for pixel classification at the end of each deep grid. You can see the picture pyramid used in the CNN system in Figure 11.

The Laplacian pyramid is also used to segment medical images semantically, and the reader can refer to the article in [113]. Authors have a multiresolution redevelopment architecture based on a Laplacian pyramid that utilizes higher-resolution map jump connections and a polygonal gate to change reconstructed boundaries with low-resolution function maps gradually. Reference [114] introduces a method of scene interpretation, and, through the Laplacian pyramid, the raw input image is transformed. In comparison, CNN creates a series of feature charts and generates each scale in two phases.

5.2. Atrous Spatial Pyramid Pooling (ASPP)

Inspired by the image pyramid technique, [92] Atrous Spatial Pyramid Pooling (ASPP) is suggested to be done to provide robust object segmentation at different scales. ASPP explores the powerful fields-of-views (FOV) and the convolution feature layer with a multisampling rate filter and then captures the artifacts at various scales in the scene. ASPP architecture is seen in Figure 12.

5.3. Pyramid Pooling

According to the Pyramid Pooling shown in Figure 11, through gathering image data based on various regions, [115] exploits global knowledge capacity and calls its pyramid scene parsing network, known as PSPNet. The excellent results therein show that, with pyramidal pooling, a PSPNet brings a new mIoU score record of 85.4 percent in PASCAL VOC 2012 and brings 80.2 percent in Cityscapes dataset by experiments in [115]. Pyramid Pooling adopts multiple pooling size scales and applies the output to the original size for sampling processing. Finally, to shape a composite feature profile, it obtains the findings. Different scales of the size of the pool with different colors are marked in Figure 13. In general, pyramidal pooling can be used for any mapping of features. For example, the program in [115] applies pyramidal pooling in the pool5 layer.

5.4. Feature Pyramid

As mentioned in research backgrounds such as [117], the feature pyramid is an important component in image work for recognizing objects of various sizes. Object detectors have avoided displaying pyramids with recent DL methods because the computational volume and memory are compact. In [117], authors use the CNN multiscale pyramid hierarchy to build special pyramids at an additional cost. Also, a Feature Pyramid Network (FPN) has been created to construct high-level semantic maps at all scales. Machine learning also has many applications in the optimal selection of feature extraction [118–122].

6. Multilevel and Multistep Feature Methods

CNN may be known as a feature extractor [123], and, as a feature, CNN-based detection algorithms usually use the last-layer output. For dense forecasting, however, the data in this layer is too big. Instead, in localization, the primary layers can be correct but do not present the meaningful state. They describe hypercolumns as the activation vectors of all CNN units above that pixel to achieve both. You will see the form adopted as superstores in [123] in Figure 14. The FCN [86] has already approved jumps, as seen in Figure 5. The multilevel approach appears to have been used in the study [123], and multimodeling is a group approach to visualization ([125, 126]). In comparison to the multilevel approach, the multistage technique is utilized in semantic segmentation [107] to improve its accuracy and speed, which recommends the deep layer cascade (LC) method. The deep layer cascade (LC) method consists of multiple independent models, unlike the conventional model cascade (MC) ([125, 126]). As a multisubset model cascade, the LC system uses a single deep model, classifying several basic parts into the shallow stage and concentrating the deeper stage on several hard sections. This not only increases the productivity of segmentation but also accelerates both deep network training and research (Figure 12).

7. The Most Practical Deep Learning Methods in Medical Image Segmentation

The application of DL techniques to segmentation methods and MRI imaging aiming at brain tumors has been studied extensively. In general, different structures of a convolutional neural network can be acknowledged as the best DL technique in this research with studies on a convolutional neural network with a deeper layer [127], two-way convolutional neural network [128], cascaded CNN [129], multidimensional convolutional neural network [130], fully convolutional neural network (FCNN) for training with CRF [131], three-dimensional model of the convolutional neural network, two-dimensional model of the convolutional neural network [132], extreme learning machine (ELM) [133], Growing Deep Convolutional Network (GCNN) [134], complete convolutional neural network with Atheros convolution pyramid features [135], three-dimensional convolutional neural network test-time augmentation [136], and convolutional neural network referred to as CRF-based multicascade [137]. All the weaknesses along with the general application of all the advantages of the classical methods are presented in Table 2, and it is seen in the available and studied methods and also all of them have a wide range of applications. Yang et al. have used a portable evanescent wave sensor to detect SARS-CoV-2 using a CRISPR-based [138]. Reference [139] also uses DL methods to detect and classify breast tumors. Three different DL architectures, GoogLeNet, VGGNet, and ResNet, have been considered, and analysis has been performed between these methods. Visual detection and evaluation of breast tumors with DL principles are also presented in [140], which uses the combined methods of K-means and SURF algorithms in the structure of DL networks based on multiclass support vector machine. The detection of breast cancer using an extreme learning machine (ELM) based on feature fusion with deep convolutional neural network features is presented in [141]. Also, in [142], the extraction of a distinct pattern for the histopathological image classification of breast cancer has been done through an automated structure based on a convolutional neural network. All the weaknesses along with the general application of all the advantages of the classical methods are presented in Table 3; and it is seen in the available and studied methods and all of them have a high range of applications. Figure 15 shows chest radiographs in healthy individuals and COVID-19 patients, respectively (see Table 4).

(a)

(b)

8. Discussion

A primary application in image processing and computer vision is image medical semantic segmentation. In addition to a brief overview of image semantic segmentation and traditional medicine, this article discusses recent advances in image semantic segmentation, particularly based on deep convolutional neural networks in the following aspects: (1) fully convolutional network, (2) method of sampling method, (3) combining FCN with CRF methods, (4) dilated convolution approaches, (5) progress in the main pillar of networks, (6) pyramid methods, and (7) multistage properties and multilevel methods. So far, there have been more and better ways to segment medical images semantically more accurately or faster or both with higher accuracy and speed, as well as better performance. Finally, the authors of this article hope that this review of recent advances in image medical semantic segmentation will help researchers in this field.

Maghdid et al. [21] reviewed a comprehensive, preprocessed dataset on X-rays and CT scan images from a variety of sources and provided an algorithm for accurate diagnosis of COVID-19 using DL and transmission learning tools. Also, a modified model was used by CNN and AlexNet as a pretested network on ready-made X-ray and CT scan datasets. After extensive experiments in both datasets, it has been shown that the proposed COVID-19 model predicts high accuracy and low response time. It is important to note that their proposed DL pattern has shown equivalent performance compared to that of a specialist radiologist. In addition, it can significantly improve the efficiency of radiologists while performing clinical practice [21]. Researchers are searching for new ways of screening, and the DL added to the chest X-rays of patients has shown positive outcomes. The computational cost of these approaches is still high considering their popularity, which causes difficulties with accessing them. Therefore the main purpose of this research is to accommodate the COVID-19 screening issue in chest X-ray with a reliable and successful approach in terms of memory and processing time. DL is a branch of artificial intelligence (AI) machine learning related to algorithms that are inspired as artificial neural networks by the structure and operation of the brain by using far higher-quality input images without any processing time. In addition, it is faster and cheaper to embed these versions in devices with more limited settings such as smartphones. To make use of embedded and large-scale devices, models can need little memory and carry out research quickly; and it encourages smartphones and emergency devices to work with them. DL models are complicated, so, to avoid inserting connections, a large number of things are necessary. In the training suite, for example, where the learning network performs well, to have less performance in the test suite, a large number of items are required. Unfortunately, there is not much data available for most real-world problems, even though the dataset is still small. Efficient training in deep neural networks has also been rendered possible through studying data transmission and amplification strategies in the small number of COVID-19-related images [144]. A popular approach for survival analysis and event prediction is the CPH model. This is therefore a semiparametric model, which suggests that the probability of misdiagnosis is a linear mixture of the clinical variables of the patient. In a fully data-driven way, the DL model can learn and infer high-order nonlinear interactions between clinical variables and disease effects. Data improvement techniques in DL will also make the model more robust to information noise and lost information, which usually happens in clinical datasets. It is also possible to expand the DL model to incorporate time-dependent factors such as vital signs and elevated visual attributes such as CT or X-ray images. It is inevitable to lose data on certain factors in reality and the real world. Data lost in less than three variables was then permitted in authors’ online measurement tool, and risk evaluation based on DL methods can still be given by the field. In the clinical experience of Liang et al. [152], mild cases of COVID-19 are generally limited, and these are acute cases that need to be further investigated by physicians. Classified cases of their patients are clinically and economically expensive to manage COVID-19, especially due to the rapid outbreak of the disease which can happen and the high mortality rate related to acute disease, which has a high cost. By submitting clinical information online, medical personnel can use the predicted risk index to hospitalize patients and accordingly arrange patient treatment plans. In this way, medical resources can be appropriately allocated [152]. Arora et al. [26] suggested a DL model for estimating the number of patients who may have COVID-19 infection. They estimated the number of new cases of new coronaviruses in various states of the Indian Union for a span of one day to one week. For prediction, they utilized repetitive neural networks and Long Short-Term Memory (LSTM) and then tested several LSTM models in the Indian dataset and concluded that deeper LSTM models such as stacked LSTM, circular LSTM, and two-way LSTM were more accurate than simple LSTM models. To date, no research studies on COVID-19 cases have been reported from all Indian states, according to the authors. In one study, to predict the number of COVID-19-positive cases in the Indian states, Arora and colleagues suggested DL models. Because of the growing number of positive cases in India, exploratory data analysis has been undertaken. Depending on the number of cases and the daily growth rate, the government classifies states into mild, moderate, and severe areas to take strong action against the quarantine of the entire country, and this may cause economic and social problems. As predictive models, recurrent neural networks (RNN) are used based on long-term and short-term memory (LSTM) cells. In 32 states/union states, LSTM types such as deep LSTM, circular LSTM, and bidirectional LSTM models have been tested, and the model is chosen with maximum accuracy based on absolute error. Based on estimation errors, the best outcome is the two-way LSTM, and the worst result is the hanging LSTM [26]. Ardakani et al. [149] suggested a DL-based CAD method to classify COVID-19 versus other pneumonic and abnormal pneumonia in research. They proposed that the DL method will assist radiologists to diagnose the disease associated with COVID-19, and they used ten convolutional neural networks (CNNs) to identify COVID-19-related diseases. Ten well-known CNNs were used in this analysis to provide a detailed view of the role of artificial intelligence in COVID-19 diagnosis. Data have shown that DL can distinguish COVID-19 with high accuracy from other pneumonia and viral diseases. For the ResNet-101 and Xeption networks, the best findings have been found. In the classification of COVID-19 and non-COVID-19 diseases, however, the Xeption network was most successful, but it did not have the highest sensitivity. In comparison, ResNet-101 was able to detect COVID-19 infection with the highest sensitivity and present fewer features compared to the Xeption network. In diagnosing patients with COVID-19, the trend is to incorporate a system with the greatest sensitivity. The advancement of DL programs helps researchers to do fast and deep X-ray scan analysis. DL is a mixture of methods of machine learning that focuses primarily on the automated extraction and classification of image characteristics, while its applications are commonly employed in medical work, medical detection, and classification. Machine learning and DL in the application of artificial intelligence for mining, pattern analysis from data, have been created as a discipline. To further evaluate the deep cognition method, Apostolopoulos et al. performed an experiment using six common lung diseases, including COVID-19. In this method, its capabilities in differentiating between different diseases are evaluated. Fine-tuning a deep network, in the context of DL, is a common approach for both learning the properties of depth and maintaining the method for extracting global properties, which exist in each image as different shapes. Specific research to detect potential trademarks focuses on X-ray images, and these biomarkers can be substantially correlated with COVID-19 disease. However, DL derives from images a large range of high-dimensional features, and some of these features may be known as real image markers. Li et al. studied the effect of self-assembly on fluorescent in magnetic fluid flow and its use for a new COVID-19 detection [149]. Recently, many studies have been done on various subjects about COVID-19, such as scheduling problems [153], climate change [154], sunspot assessment [155], disease severity and industry [156], energy after COVID-19 pandemic, travel-related risks among pandemics [157], and predictive modeling [148].

9. Limitations

The limitations of this research are mentioned in several aspects. First, the CT validation dataset is collected at one center, which may not represent all COVID-19 patients in other geographic areas. The generalization of the DL system must be approved in several centers. Second, the system is designed to determine the outbreak of the disease and may not be effective in measuring other pneumonia, such as bacterial pneumonia. Finally, in the next work, the authors will develop a system for quantifying the total intensity of pneumonia using transfer learning. Not ready for production, the researchers hope that the results obtained by COVID-Net in the COVIDX test dataset will be available as open source with descriptions of the open-source dataset. CXR images were used to accelerate the development of high-precision DL solutions for the diagnosis of COVID-19 patients. The future work will continue to include increasing accuracy and PPV for COVID-19 with the collection of new data, as well as the development of COVID-Net for risk classification for survival analysis, patient status prediction, and length of hospital stay.

10. Conclusion

In this study, the potential of deep learning methods in COVID-19 diagnosis is investigated. This study has reviewed the classification systems based on DL to assess the extent of the disease. This system not only automatically contours the infected areas but also measures their shape, volume, and percentage of infection on a CT scan of patients with COVID-19. The methods involve radiologists to intervene effectively in the results of DL segmentation and repeatedly add more tutorials to update the model, thus accelerating the algorithm’s development cycle. CT imaging has become an effective tool for screening patients with COVID-19 and for assessing COVID-19 levels. However, radiologists do not thave a computer tool to accurately determine the severity of COVID-19, for example, the percentage of infection in the lungs. DL has become a common method in medical image analysis and has been used in the analysis of lung diseases. Using this deep learning automated segmentation, many studies on imaging quantification and its association with syndromes, epidemiology, and therapeutic responses can provide further information on improving the diagnosis and treatment of COVID-19. An AI algorithm can be created quickly from one or more algorithms that do the same thing. This is in contrast to the standard method for generating a DL algorithm, which requires several steps. In order to review the data, expert annotations are needed at the data collection point at which a large number of samples need to be taken. The second is the process of training in which the data obtained is used to train network models. Every category should be well represented so that the training can be generalized during the test process to the new objects found by the network. A great number of network parameters (typically in the order of millions) are created automatically in this learning process. The third step is the experiment in which the network is presented with another collection of objects not included in the testing and the network performance is statistically evaluated to determine its classification. There is no solution that fits all; we hope that the positive results obtained by COVID-Net will be present in the COVIDX test dataset. Images are used to boost the advancement of highly accurate DL solutions for the diagnosis of COVID-19 patients and accelerate the treatment of patients. Future pathways including continuing to enhance sensitivity and PPV to COVID-19 disease by collecting new data as well as extending the suggested COVID-Net to risk classification for analysis, patient status prediction, and length of hospital stay will be useful. [158–164]

Data Availability

This is a review paper and data sharing is not applicable.

Conflicts of Interest

The authors declare no conflicts of interest.

Acknowledgments

The funding sources had no involvement support in the study design, collection, analysis, or interpretation of data, in writing of the manuscript, or in the decision to submit the manuscript for publication.

References

M. Thoma, “A survey of semantic segmentation,” 2016, http://arxiv.org/abs/1602.06541.
View at: Google Scholar
M. Rezaei, F. Farahanipad, A. Dillhoff, R. Elmasri, and V. Athitsos, “Weakly-supervised hand part segmentation from depth images,” in Proceedings of the 14th PErvasive Technologies Related to Assistive Environments Conference, pp. 218–225, Corfu, Greece, June 2021.
View at: Publisher Site | Google Scholar
Y. Li, H. Qi, J. Dai, X. Ji, and Y. Wei, “Fully convolutional instance-aware semantic segmentation,” in Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, pp. 4438–4446, Honolulu, HI, USA, July 2017.
View at: Publisher Site | Google Scholar
S. Maldonado-Bascon, S. Lafuente-Arroyo, P. Gil-Jimenez, H. Gomez-Moreno, and F. Lopez-Ferreras, “Road-sign detection and recognition based on support vector machines,” IEEE Transactions on Intelligent Transportation Systems, vol. 8, no. 2, pp. 264–278, 2007.
View at: Publisher Site | Google Scholar
A. Cohen, E. Rivlin, I. Shimshoni, and E. Sabo, “Memory based active contour algorithm using pixel-level classified images for colon crypt segmentation,” Computerized Medical Imaging and Graphics, vol. 43, pp. 150–164, 2015.
View at: Publisher Site | Google Scholar
C. Huang, L. S. Davis, and J. R. G. Townshend, “An assessment of support vector machines for land cover classification,” International Journal of Remote Sensing, vol. 23, no. 4, pp. 725–749, 2002.
View at: Publisher Site | Google Scholar
N. Moon, E. Bullitt, K. van Leemput, and G. Gerig, “Automatic brain and tumor segmentation,” in Proceedings of the Medical Image Computing and Computer-Assisted Intervention-MICCAI 2002, pp. 372–379, Springer Berlin Heidelberg, Tokyo, Japan, September 2002.
View at: Publisher Site | Google Scholar
G.-Q. Wei, K. Arbter, and G. Hirzinger, “Automatic tracking of laparoscopic instruments by color coding,” in Proceedings of the International Conference on Computer Vision, Virtual Reality, and Robotics in Medicine, pp. 357–366, Springer Berlin Heidelberg, Grenoble, France, March 1997.
View at: Publisher Site | Google Scholar
H.-L. Chen, G. Wang, C. Ma, Z.-N. Cai, W.-B. Liu, and S.-J. Wang, “An efficient hybrid kernel extreme learning machine approach for early diagnosis of Parkinson’s disease,” Neurocomputing, vol. 184, pp. 131–144, 2016.
View at: Publisher Site | Google Scholar
W. Zhou, L. Yu, Y. Zhou, W. Qiu, M.-W. Wu, and T. Luo, “Local and global feature learning for blind quality evaluation of screen content and natural scene images,” IEEE Transactions on Image Processing, vol. 27, no. 5, pp. 2086–2095, 2018.
View at: Publisher Site | Google Scholar
Z. Lv, L. Qiao, A. K. Singh, and Q. Wang, “Fine-grained visual computing based on deep learning,” ACM Transactions on Multimedia Computing, Communications, and Applications, vol. 17, no. 1s, pp. 1–19, 2021.
View at: Publisher Site | Google Scholar
R. Liu, X. Wang, H. Lu et al., “SCCGAN: style and characters inpainting based on CGAN,” Mobile Networks and Applications, vol. 26, no. 1, pp. 3–12, 2021.
View at: Publisher Site | Google Scholar
C. Li, L. Hou, B. Y. Sharma et al., “Developing a new intelligent system for the diagnosis of tuberculous pleural effusion,” Computer Methods and Programs in Biomedicine, vol. 153, pp. 211–225, 2018.
View at: Publisher Site | Google Scholar
L. Wang and A. Wong, “COVID-Net: A tailored deep convolutional neural network design for detection of COVID-19 cases from chest X-ray images,” 2020, http://arxiv.org/abs/2003.09871.
View at: Google Scholar
M. Zhang, Y. Chen, and W. Susilo, “PPO-CPQ: a privacy-preserving optimization of clinical pathway query for e-healthcare systems,” IEEE Internet of Things Journal, vol. 7, no. 10, pp. 10660–10672, 2020.
View at: Publisher Site | Google Scholar
A. Ala, F. E. Alsaadi, M. Ahmadi, and S. Mirjalili, “Optimization of an appointment scheduling problem for healthcare systems based on the quality of fairness service using whale optimization algorithm and NSGA-II,” Scientific Reports, vol. 11, no. 1, pp. 19816–19819, 2021.
View at: Publisher Site | Google Scholar
Q. Xu, Q. Guo, C.-X. Wang et al., “Network differentiation: a computational method of pathogenesis diagnosis in traditional Chinese medicine based on systems science,” Artificial Intelligence in Medicine, vol. 118, Article ID 102134, 2021.
View at: Publisher Site | Google Scholar
J. Deng, W. Dong, R. Socher, L.-J. Li, K. Kai Li, and L. Fei-Fei, “ImageNet: a large-scale hierarchical image database,” in Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition, pp. 248–255, Miami, FL, USA, June 2009.
View at: Publisher Site | Google Scholar
M. Ahmadi, T. Ali, D. Javaheri, A. Masoumian, S. Jafarzadeh Ghoushchi, and Y. Pourasad, “DQRE-SCnet: a novel hybrid approach for selecting users in federated learning with deep-Q-reinforcement learning based on spectral clustering,” Journal of King Saud University-Computer and Information Sciences, 2021.
View at: Publisher Site | Google Scholar
E. E.-D. Hemdan, M. A. Shouman, and M. E. Karar, “COVIDX-Net: a framework of deep learning classifiers to diagnose COVID-19 in X-ray images,” 2020, http://arxiv.org/abs/2003.11055.
View at: Google Scholar
H. S. Maghdid, A. T. Asaad, K. Z. Ghafoor, A. S. Sadiq, and M. K. Khan, “Diagnosing COVID-19 pneumonia from X-ray and CT images using deep learning and transfer learning algorithms,” 2020, http://arxiv.org/abs/2004.00038.
View at: Google Scholar
D. Martin and C. Fowlkes, The Berkeley Segmentation Database and Benchmark, University of California, Berkeley, CA, USA, 2001, http://www.eecs.berkeley.edu/Research/Projects/CS/vision/bsds.
B. C. Russell, A. Torralba, K. P. Murphy, and W. T. Freeman, “LabelMe: a database and web-based tool for image annotation,” International Journal of Computer Vision, vol. 77, no. 1-3, pp. 157–173, 2008.
View at: Publisher Site | Google Scholar
S. K. Hasan and C. A. Linte, “A modified U-Net convolutional network featuring a nearest-neighbor re-sampling-based elastic-transformation for brain tissue characterization and segmentation,” in Proceedings of the 2018 IEEE Western New York Image and Signal Processing Workshop (WNYISPW), pp. 1–5, IEEE, Rochester, NY, USA, October 2018.
View at: Publisher Site | Google Scholar
S. Charan, M. J. Khan, and K. Khurshid, “Breast cancer detection in mammograms using convolutional neural network,” in Proceedings of the 2018 International Conference on Computing, Mathematics and Engineering Technologies (iCoMET), pp. 1–5, IEEE, Sukkur, Sindh, Pakistan, March 2018.
View at: Publisher Site | Google Scholar
P. Arora, H. Kumar, and B. K. Panigrahi, “Prediction and analysis of COVID-19 positive cases using deep learning models: a descriptive case study of India,” Chaos, Solitons & Fractals, vol. 139, Article ID 110017, 2020.
View at: Publisher Site | Google Scholar
J. Xia, H. Chen, Q. Li et al., “Ultrasound-based differentiation of malignant and benign thyroid Nodules: an extreme learning machine approach,” Computer Methods and Programs in Biomedicine, vol. 147, pp. 37–49, 2017.
View at: Publisher Site | Google Scholar
L. Bourdev, S. Maji, T. Brox, and J. Malik, “Detecting people using mutually consistent poselet activations,” in Proceedings of the Computer Vision-ECCV 2010, pp. 168–181, Springer Berlin Heidelberg, Heraklion, Crete, Greece, September 2010.
View at: Publisher Site | Google Scholar
D. G. Lowe, “Distinctive image features from scale-invariant keypoints,” International Journal of Computer Vision, vol. 60, no. 2, pp. 91–110, 2004.
View at: Publisher Site | Google Scholar
D. C. He and L. Wang, “Texture unit, texture spectrum, and texture analysis,” IEEE Transactions on Geoscience and Remote Sensing, vol. 28, no. 4, pp. 509–512, 1990.
View at: Publisher Site | Google Scholar
H. Bay, A. Ess, T. Tuytelaars, and L. Van Gool, “Speeded-up robust features (SURF),” Computer Vision and Image Understanding, vol. 110, no. 3, pp. 346–359, 2008.
View at: Publisher Site | Google Scholar
K. G. Derpanis, The Harris Corner Detector, York University, Toronto, Canada, 2004.
J. Shi and C. Tomasi, “Good features to track,” in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 593–600, Seattle, WA, USA, June 1994.
View at: Publisher Site | Google Scholar
G. Medioni and Y. Yasumoto, “Corner detection and curve representation using cubic B-splines,” Computer Vision, Graphics, and Image Processing, vol. 39, no. 3, pp. 267–278, 1987.
View at: Publisher Site | Google Scholar
S. M. Smith and J. M. Brady, “Susan-a new approach to low level image processing,” International Journal of Computer Vision, vol. 23, no. 1, pp. 45–78, 1997.
View at: Publisher Site | Google Scholar
E. Rosten and T. Drummond, “Fusing points and lines for high performance tracking,” Tenth IEEE International Conference on Computer Vision (ICCV’05) Volume 1, vol. II, pp. 1508–1515, 2005.
View at: Publisher Site | Google Scholar
E. Rosten, R. Porter, and T. Drummond, “Faster and better: a machine learning approach to corner detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 32, no. 1, pp. 105–119, 2010.
View at: Publisher Site | Google Scholar
E. Mair, G. D. Hager, D. Burschka, M. Suppa, and G. Hirzinger, “Adaptive and generic corner detection based on the accelerated segment test,” in Proceedings of the Computer Vision-ECCV 2010, pp. 183–196, Springer Berlin Heidelberg, Heraklion, Crete, Greece, September 2010.
View at: Publisher Site | Google Scholar
S. Leutenegger, M. Chli, and R. Y. Siegwart, “BRISK: binary Robust invariant scalable keypoints,” in Proceedings of the 2011 International Conference on Computer Vision, pp. 2548–2555, Barcelona, Spain, November 2011.
View at: Publisher Site | Google Scholar
C. S. Venegas-Barrera and J. Manjarrez, “Visual categorization with bags of keypoints,” Revista Mexicana de Biodiversidad, vol. 82, no. 1, pp. 179–191, 2011.
View at: Publisher Site | Google Scholar
T. Brox, L. Bourdev, S. Maji, and J. Malik, “Object segmentation by alignment of poselet activations to image contours,” in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 2225–2232, Colorado Springs, CO, USA, June 2011.
View at: Publisher Site | Google Scholar
S.-C. Zhu, C.-E. Guo, Y. Wang, and Z. Xu, “What are Textons?” International Journal of Computer Vision, vol. 62, no. 1, pp. 121–143, 2005.
View at: Publisher Site | Google Scholar
F. Liu, G. Zhang, and J. Lu, “Multi-source heterogeneous unsupervised domain adaptation via fuzzy-relation neural networks,” IEEE Transactions on Fuzzy Systems, vol. 29, 2020.
View at: Publisher Site | Google Scholar
Q. Xu, Y. Zeng, W. Tang et al., “Multi-task joint learning model for segmenting and classifying tongue images using a deep neural network,” IEEE Journal of Biomedical and Health Informatics, vol. 24, no. 9, pp. 2481–2489, 2020.
View at: Publisher Site | Google Scholar
H. Liu, J. Liu, S. Hou, T. Tao, and J. Han, “Perception consistency ultrasound image super-resolution via self-supervised CycleGAN,” Neural Computing and Applications, pp. 1–11, 2021.
View at: Publisher Site | Google Scholar
X. Zhao, X. Zhang, Z. Cai et al., “Chaos enhanced grey wolf optimization wrapped ELM for diagnosis of paraquat-poisoned patients,” Computational Biology and Chemistry, vol. 78, pp. 481–490, 2019.
View at: Publisher Site | Google Scholar
D. J. Strauss and J. A. Hartigan, “Clustering algorithms,” Biometrics, vol. 31, no. 3, p. 793, 1975.
View at: Publisher Site | Google Scholar
H. Mobahi, S. R. Rao, A. Y. Yang, S. S. Sastry, and Y. Ma, “Segmentation of natural images by texture and boundary compression,” International Journal of Computer Vision, vol. 95, no. 1, pp. 86–98, 2011.
View at: Publisher Site | Google Scholar
R. Kimmel and A. M. Bruckstein, “Regularized Laplacian zero crossings as optimal edge integrators,” International Journal of Computer Vision, vol. 53, no. 3, pp. 225–243, 2003.
View at: Publisher Site | Google Scholar
H. Yu, W. Li, C. Chen et al., “Dynamic gaussian bare-bones fruit fly optimizers with abandonment mechanism: method and analysis,” Engineering with Computers, 2020.
View at: Google Scholar
X. Zhao, D. Li, B. Yang, C. Ma, Y. Zhu, and H. Chen, “Feature selection based on improved ant colony optimization for online detection of foreign fiber in cotton,” Applied Soft Computing, vol. 24, pp. 585–596, 2014.
View at: Publisher Site | Google Scholar
C. Yu, M. Chen, K. Cheng et al., “SGOA: annealing-behaved grasshopper optimizer for global tasks,” Engineering with Computers, pp. 1–28, 2021.
View at: Publisher Site | Google Scholar
T. Lindeberg and M.-X. Li, “Segmentation and classification of edges using minimum description length Approximation and complementary junction cues,” Computer Vision and Image Understanding, vol. 67, no. 1, pp. 88–98, 1997.
View at: Publisher Site | Google Scholar
L. Shen, H. Chen, Z. Yu et al., “Evolving support vector machines using fruit fly optimization for medical data classification,” Knowledge-Based Systems, vol. 96, pp. 61–75, 2016.
View at: Publisher Site | Google Scholar
W. Shan, Z. Qiao, A. Asghar Heidari, H. Chen, H. Turabieh, and Y. Teng, “Double adaptive weights for stabilization of moth flame optimizer: balance analysis, engineering cases, and medical diagnosis,” Knowledge-Based Systems, vol. 214, Article ID 106728, 2020.
View at: Publisher Site | Google Scholar
J. Shotton, J. Winn, C. Rother, and A. Criminisi, “TextonBoost: joint appearance, shape and context modeling for multi-class object recognition and segmentation,” in Proceedings of the Computer Vision-ECCV 2006, pp. 1–15, Springer Berlin Heidelberg, Graz, Austria, May 2006.
View at: Google Scholar
A. M. N. Darwiesh, N. M. A.-E. Maboud, A. M. R. Khalil, and A. M. ElSharkawy, “Role of magnetic resonance spectroscopy & diffusion weighted imaging in differentiation of supratentorial brain tumors,” The Egyptian Journal of Radiology and Nuclear Medicine, vol. 47, no. 3, pp. 1037–1042, 2016.
View at: Publisher Site | Google Scholar
A. Aslam, E. Khan, and M. M. S. Beg, “Improved edge detection algorithm for brain tumor segmentation,” Procedia Computer Science, vol. 58, pp. 430–437, 2015.
View at: Publisher Site | Google Scholar
W. Qiao, M. Khishe, and S. Ravakhah, “Underwater targets classification using local wavelet acoustic pattern and Multi-layer perceptron neural network optimized by modified whale optimization algorithm,” Ocean Engineering, vol. 219, Article ID 108415, 2021.
View at: Publisher Site | Google Scholar
Q. Ain, M. A. Jaffar, and T.-S. Choi, “Fuzzy anisotropic diffusion based segmentation and texture based ensemble classification of brain tumor,” Applied Soft Computing, vol. 21, pp. 330–340, 2014.
View at: Publisher Site | Google Scholar
S. Hassantabar, M. Ahmadi, and A. Sharifi, “Diagnosis and detection of infected tissue of COVID-19 patients based on lung X-ray image using convolutional neural network approaches,” Chaos, Solitons & Fractals, vol. 140, Article ID 110170, 2020.
View at: Publisher Site | Google Scholar
S. Dorosti, S. Jafarzadeh Ghoushchi, E. Sobhrakhshankhah, M. Ahmadi, and A. Sharifi, “Application of gene expression programming and sensitivity analyses in analyzing effective parameters in gastric cancer tumor size and location,” Soft Computing, vol. 24, no. 13, pp. 9943–9964, 2020.
View at: Publisher Site | Google Scholar
A. Sharifi, M. Ahmadi, M. A. Mehni, S. Jafarzadeh Ghoushchi, and Y. Pourasad, “Experimental and numerical diagnosis of fatigue foot using convolutional neural network,” Computer Methods in Biomechanics and Biomedical Engineering, pp. 1–13, 2021.
View at: Publisher Site | Google Scholar
I. Laradji, P. Rodriguez, O. Manas et al., “A weakly supervised consistency-based learning method for covid-19 segmentation in ct images,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp. 2453–2462, Waikoloa, HI, USA, 2021.
View at: Publisher Site | Google Scholar
M. Karnan and K. Selvanayaki, “Improved implementation of brain MR image segmentation using meta heuristic algorithms,” in Proceedings of the 2010 IEEE International Conference on Computational Intelligence and Computing Research, pp. 702–705, Coimbatore, India, December 2010.
View at: Publisher Site | Google Scholar
P. Ghosh, K. Mali, and S. K. Das, “Chaotic firefly algorithm-based fuzzy C-means algorithm for segmentation of brain tissues in magnetic resonance images,” Journal of Visual Communication and Image Representation, vol. 54, pp. 63–79, 2018.
View at: Publisher Site | Google Scholar
Q. Zhu, B. Du, P. Yan, H. Lu, and L. Zhang, “Shape prior constrained PSO model for bladder wall MRI segmentation,” Neurocomputing, vol. 294, pp. 19–28, 2018.
View at: Publisher Site | Google Scholar
S. Alagarsamy, K. Kamatchi, V. Govindaraj, Y.-D. Zhang, and A. Thiyagarajan, “Multi-channeled MR brain image segmentation: a new automated approach combining BAT and clustering technique for better identification of heterogeneous tumors,” Biocybernetics and Biomedical Engineering, vol. 39, no. 4, pp. 1005–1035, 2019.
View at: Publisher Site | Google Scholar
R. Rouhi, M. Jafari, S. Kasaei, and P. Keshavarzian, “Benign and malignant breast tumors classification based on region growing and CNN segmentation,” Expert Systems with Applications, vol. 42, no. 3, pp. 990–1002, 2015.
View at: Publisher Site | Google Scholar
S. Kaymak, A. Helwan, and D. Uzun, “Breast cancer image classification using artificial neural networks,” Procedia Computer Science, vol. 120, pp. 126–131, 2017.
View at: Publisher Site | Google Scholar
M. Karabatak, “A new classifier for breast cancer detection based on Naïve Bayesian,” Measurement, vol. 72, pp. 32–36, 2015.
View at: Publisher Site | Google Scholar
F. Wang, S. Zhang, and L. M. Henderson, “Adaptive decision-making of breast cancer mammography screening: a heuristic-based regression model,” Omega, vol. 76, pp. 70–84, 2018.
View at: Publisher Site | Google Scholar
D. C. Pereira, R. P. Ramos, and M. Z. do Nascimento, “Segmentation and detection of breast cancer in mammograms combining wavelet analysis and genetic algorithm,” Computer Methods and Programs in Biomedicine, vol. 114, no. 1, pp. 88–101, 2014.
View at: Publisher Site | Google Scholar
L. Xu, S. Jiang, J. Wu, and Q. Zou, “An in silico approach to identification, categorization and prediction of nucleic acid binding proteins,” Briefings in Bioinformatics, vol. 22, no. 3, Article ID bbaa171, 2021.
View at: Publisher Site | Google Scholar
F. R. Cordeiro, W. P. Santos, and A. G. Silva-Filho, “An adaptive semi-supervised fuzzy growcut algorithm to segment masses of regions of interest of mammographic images,” Applied Soft Computing, vol. 46, pp. 613–628, 2016.
View at: Publisher Site | Google Scholar
L. Ahmed, M. M. Iqbal, H. Aldabbas, S. Khalid, Y. Saleem, and S. Saeed, “Images data practices for semantic segmentation of breast cancer using deep neural network,” Journal of Ambient Intelligence and Humanized Computing, pp. 1–7, 2020.
View at: Publisher Site | Google Scholar
H. Lee, J. Park, and J. Y. Hwang, “Channel attention module with multiscale grid average pooling for breast cancer segmentation in an ultrasound image,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 67, no. 7, pp. 1344–1353, 2020.
View at: Publisher Site | Google Scholar
K. B. Soulami, N. Kaabouch, M. N. Saidi, and A. Tamtaoui, “Breast cancer: one-stage automated detection, segmentation, and classification of digital mammograms using UNet model based-semantic segmentation,” Biomedical Signal Processing and Control, vol. 66, Article ID 102481, 2021.
View at: Publisher Site | Google Scholar
K. Huang, Y. Zhang, H. D. Cheng, P. Xing, and B. Zhang, “Semantic segmentation of breast ultrasound image with fuzzy deep learning network and breast anatomy constraints,” Neurocomputing, vol. 450, pp. 319–335, 2021.
View at: Publisher Site | Google Scholar
R. M. S. d. Oliveira, R. C. F. Araújo, F. J. B. Barros et al., “A system based on artificial neural networks for automatic classification of hydro-generator stator windings partial discharges,” Journal of Microwaves, Optoelectronics and Electromagnetic Applications, vol. 16, no. 3, pp. 628–645, 2017.
View at: Publisher Site | Google Scholar
Y. Bengio, “Learning deep architectures for AI,” Foundations and Trends in Machine Learning, vol. 2, no. 1, pp. 1–127, 2009.
View at: Publisher Site | Google Scholar
H. Larochelle and Y. Bengio, “Classification using discriminative restricted Boltzmann machines,” in Proceedings of the 25th International Conference on Machine Learning-ICML’08, pp. 536–543, Helsinki, Finland, July 2008.
View at: Publisher Site | Google Scholar
M. Ahmadi, A. Sharifi, and S. Khalili, “Presentation of a developed sub-epidemic model for estimation of the COVID-19 pandemic and assessment of travel-related risks in Iran,” Environmental Science and Pollution Research, vol. 28, no. 12, pp. 14521–14529, 2021.
View at: Publisher Site | Google Scholar
S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997.
View at: Publisher Site | Google Scholar
J. L. Olazagoitia, J. A. Perez, and F. Badea, “Identification of tire model parameters with artificial neural networks,” Applied Sciences, vol. 10, no. 24, p. 9110, 2020.
View at: Publisher Site | Google Scholar
E. Shelhamer, J. Long, and T. Darrell, “Fully convolutional networks for semantic segmentation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, no. 4, pp. 640–651, 2017.
View at: Publisher Site | Google Scholar
M. Soltaninejad, L. Zhang, T. Lambrou, N. Allinson, and X. Ye, “Multimodal MRI brain tumor segmentation using random forests with features learned from fully convolutional neural network,” 2017, http://arxiv.org/abs/1704.08134.
View at: Google Scholar
H. Noh, S. Hong, and B. Han, “Learning deconvolution network for semantic segmentation,” in Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV), pp. 1520–1528, Santiago, Chile, December 2015.
View at: Publisher Site | Google Scholar
R. Mohan, “Deep deconvolutional networks for scene parsing,” 2014, http://arxiv.org/abs/1411.4101.
View at: Google Scholar
J. B. De Monvel, E. Scarfone, S. Le Calvez, and M. Ulfendahl, “Image-adaptive deconvolution for three-dimensional deep biological imaging,” Biophysical Journal, vol. 85, no. 6, pp. 3991–4001, 2003.
View at: Publisher Site | Google Scholar
N. Alipour and A. Behrad, “Semantic segmentation of JPEG blocks using a deep CNN for non-aligned JPEG forgery detection and localization,” Multimedia Tools and Applications, vol. 79, no. 11-12, pp. 8249–8265, 2020.
View at: Publisher Site | Google Scholar
L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, “DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 40, no. 4, pp. 834–848, 2018.
View at: Publisher Site | Google Scholar
M. Mostajabi, P. Yadollahpour, and G. Shakhnarovich, “Feedforward semantic segmentation with zoom-out features,” in Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3376–3385, Boston, MA, USA, June 2015.
View at: Publisher Site | Google Scholar
Z. Liu, X. Li, P. Luo, C.-C. Loy, and X. Tang, “Semantic image segmentation via deep parsing network,” in Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV), pp. 1377–1385, Santiago, Chile, December 2015.
View at: Publisher Site | Google Scholar
X.-Y. Zhou, J.-Q. Zheng, and G.-Z. Yang, “Atrous convolutional neural network (ACNN) for biomedical semantic segmentation with dimensionally lossless feature maps,” 2019, http://arxiv.org/abs/1901.09203.
View at: Google Scholar
F. Yu and V. Koltun, “Multi-scale context aggregation by dilated convolutions,” 2016, http://arxiv.org/abs/1511.07122.
View at: Google Scholar
M. Holschneider, R. Kronland-Martinet, J. Morlet, and P. Tchamitchian, “A real-time algorithm for signal analysis with the help of the wavelet transform,” in Inverse Problems and Theoretical Imaging, pp. 286–297, Springer Berlin Heidelberg, Berlin, Germnay, 1990.
View at: Publisher Site | Google Scholar
F. Yu, V. Koltun, and T. Funkhouser, “Dilated residual networks,” in Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, pp. 636–644, Honolulu, HI, USA, July 2017.
View at: Publisher Site | Google Scholar
K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” 2015, http://arxiv.org/abs/1409.1556.
View at: Google Scholar
A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet classification with deep convolutional neural networks,” Communications of the ACM, vol. 60, no. 6, pp. 84–90, 2017.
View at: Publisher Site | Google Scholar
C. Szegedy, S. Ioffe, V. Vanhoucke, and A. A. Alemi, “Inception-v4, inception-ResNet and the impact of residual connections on learning,” in Proceedings of the 2017 Thirty-First AAAI Conference on Artificial Intelligence, pp. 4278–4284, San Francisco, CA, USA, February 2017.
View at: Google Scholar
K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778, Las Vegas, NV, USA, June 2016.
View at: Publisher Site | Google Scholar
Z. Wu, C. Shen, and A. van den Hengel, “High-performance semantic segmentation using very deep fully convolutional networks,” 2016, http://arxiv.org/abs/1604.04339.
View at: Google Scholar
Z. Wu, C. Shen, and A. van den Hengel, “Wider or deeper: revisiting the ResNet model for visual recognition,” Pattern Recognition, vol. 90, pp. 119–133, 2019.
View at: Publisher Site | Google Scholar
S. Xie, R. Girshick, P. Dollar, Z. Tu, and K. He, “Aggregated residual transformations for deep neural networks,” in Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, pp. 5987–5995, Honolulu, HI, USA, July 2017.
View at: Publisher Site | Google Scholar
C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna, “Rethinking the inception architecture for computer vision,” in Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2818–2826, Las Vegas, NV, USA, June 2016.
View at: Publisher Site | Google Scholar
X. Li, Z. Liu, P. Luo, C. C. Loy, and X. Tang, “Not all pixels are equal: difficulty-aware semantic segmentation via deep layer cascade,” in Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, pp. 6459–6468, Honolulu, HI, USA, July 2017.
View at: Publisher Site | Google Scholar
E. H. Adelson and P. J. Burt, “Multi-resolution splining using A pyramid image representation,” in Proceedings of the 27th Annual Technical Symposium, pp. 204–210, San Diego, CA, USA, 1983.
View at: Publisher Site | Google Scholar
S. Seferbekov, V. Iglovikov, B. Alexander, and Alexey Shvets, “Feature pyramid network for multi-class land segmentation,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, pp. 272–275, Salt Lake City, UT, USA, June 2018.
View at: Publisher Site | Google Scholar
G. Lin, C. Shen, A. Van Den Hengel, and I. Reid, “Efficient piecewise training of deep structured models for semantic segmentation,” in Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3194–3203, Las Vegas, NV, USA, June 2016.
View at: Publisher Site | Google Scholar
L.-C. Chen, Y. Yang, J. Wang, W. Xu, and A. L. Yuille, “Attention to scale: scale-aware semantic image segmentation,” in Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3640–3649, Las Vegas, NV, USA, June 2016.
View at: Publisher Site | Google Scholar
C. Chen, W. Gong, Y. Chen, and W. Li, “Object detection in remote sensing images based on a scene-contextual feature pyramid network,” Remote Sensing, vol. 11, no. 3, p. 339, 2019.
View at: Publisher Site | Google Scholar
G. Ghiasi and C. C. Fowlkes, “Laplacian pyramid reconstruction and refinement for semantic segmentation,” in Proceedings of the Computer Vision-ECCV 2016, pp. 519–534, Amsterdam, The Netherlands, October 2016.
View at: Publisher Site | Google Scholar
C. Farabet, C. Couprie, L. Najman, and Y. Lecun, “Learning hierarchical features for scene labeling,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35, no. 8, pp. 1915–1929, 2013.
View at: Publisher Site | Google Scholar
H. Zhao, J. Shi, X. Qi, X. Wang, and J. Jia, “Pyramid scene parsing network,” in Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, pp. 6230–6239, Honolulu, HI, USA, July 2017.
View at: Publisher Site | Google Scholar
J. Li, C. Wang, L. Xu, F. Wu, H. Zhang, and B. Zhang, “Multitemporal water extraction of dongting lake and poyang lake based on an automatic water extraction and dynamic monitoring framework,” Remote Sensing, vol. 13, no. 5, p. 865, 2021.
View at: Publisher Site | Google Scholar
T.-Y. Lin, P. Dollar, R. Girshick, K. He, B. Hariharan, and S. Belongie, “Feature pyramid networks for object detection,” in Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, pp. 936–944, Honolulu, HI, USA, July 2017.
View at: Publisher Site | Google Scholar
Y. Xu, H. Chen, J. Luo, Q. Zhang, S. Jiao, and X. Zhang, “Enhanced Moth-flame optimizer with mutation strategy for global optimization,” Information Sciences, vol. 492, pp. 181–203, 2019.
View at: Publisher Site | Google Scholar
X. Xu and H.-L. Chen, “Adaptive computational chemotaxis based on field in bacterial foraging optimization,” Soft Computing, vol. 18, no. 4, pp. 797–807, 2014.
View at: Publisher Site | Google Scholar
Y. Zhang, R. Liu, X. Wang, H. Chen, and C. Li, “Boosted binary Harris hawks optimizer and feature selection,” Engineering with Computers, 2020.
View at: Publisher Site | Google Scholar
Y. Zhang, Y. Zhang, R. Liu et al., “Towards augmented kernel extreme learning models for bankruptcy prediction: algorithmic behavior and comprehensive analysis,” Neurocomputing, vol. 430, 2020.
View at: Publisher Site | Google Scholar
J. Hu, H. Chen, A. A. Heidari et al., “Orthogonal learning covariance matrix for defects of grey wolf optimizer: insights, balance, diversity, and feature selection,” Knowledge-Based Systems, vol. 213, Article ID 106684, 2021.
View at: Publisher Site | Google Scholar
B. Hariharan, P. Arbelaez, R. Girshick, and J. Malik, “Hypercolumns for object segmentation and fine-grained localization,” in Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 447–456, Boston, MA, USA, June 2015.
View at: Publisher Site | Google Scholar
Z. Zhao, B. Li, J. Lu, X. Kang, and T. Liu, “One-shot phase retrieval method for interferometry using a hypercolumns convolutional neural network,” Optics Express, vol. 29, no. 11, pp. 16406–16421, 2021.
View at: Publisher Site | Google Scholar
H. Li, Z. Lin, X. Shen, J. Brandt, and G. Hua, “A convolutional neural network cascade for face detection,” in Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 5325–5334, Boston, MA, USA, June 2015.
View at: Publisher Site | Google Scholar
P. Viola and M. Jones, “Rapid object detection using a boosted cascade of simple features,” in Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 1, Kauai, HI, USA, December 2001.
View at: Publisher Site | Google Scholar
S. Pereira, A. Pinto, V. Alves, and C. A. Silva, “Brain tumor segmentation using convolutional neural networks in MRI images,” IEEE Transactions on Medical Imaging, vol. 35, no. 5, pp. 1240–1251, 2016.
View at: Publisher Site | Google Scholar
G. Urban, M. Bendszus, F. Hamprecht, and J. Kleesiek, “Multi-modal brain tumor segmentation using deep convolutional neural networks,” in Proceedings of the Miccai-Brats, pp. 31–35, Boston, MA, USA, 2014.
View at: Google Scholar
M. Havaei, A. Davy, D. Warde-Farley et al., “Brain tumor segmentation with deep neural networks,” Medical Image Analysis, vol. 35, pp. 18–31, 2017.
View at: Publisher Site | Google Scholar
K. Hu, Z. Zhang, X. Niu et al., “Retinal vessel segmentation of color fundus images using multiscale convolutional neural network with an improved cross-entropy loss function,” Neurocomputing, vol. 309, pp. 179–191, 2018.
View at: Publisher Site | Google Scholar
K. Kamnitsas, C. Ledig, V. F. J. Newcombe et al., “Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation,” Medical Image Analysis, vol. 36, pp. 61–78, 2017.
View at: Publisher Site | Google Scholar
G. Wang, W. Li, S. Ourselin, and T. Vercauteren, “Automatic brain tumor segmentation using cascaded anisotropic convolutional neural networks,” in Proceedings of the International MICCAI Brainlesion Workshop, pp. 178–190, Quebec City, Canada, 2017.
View at: Google Scholar
L. Hu, G. Hong, J. Ma, X. Wang, and H. Chen, “An efficient machine learning approach for diagnosis of paraquat-poisoned patients,” Computers in Biology and Medicine, vol. 59, pp. 116–124, 2015.
View at: Publisher Site | Google Scholar
M. Mittal, L. M. Goyal, S. Kaur, I. Kaur, A. Verma, and D. Jude Hemanth, “Deep learning based enhanced tumor segmentation approach for MR brain images,” Applied Soft Computing, vol. 78, pp. 346–354, 2019.
View at: Publisher Site | Google Scholar
Z. Zhou, Z. He, and Y. Jia, “AFPNet: a 3D fully convolutional neural network with atrous-convolution feature pyramid for brain tumor segmentation via MRI images,” Neurocomputing, vol. 402, pp. 235–244, 2020.
View at: Publisher Site | Google Scholar
G. Wang, W. Li, S. Ourselin, and T. Vercauteren, “Automatic brain tumor segmentation using cascaded anisotropic convolutional neural networks,” in Proceedings of the Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries, pp. 178–190, Athens, Greece, October 2018.
View at: Publisher Site | Google Scholar
K. Hu, Q. Gan, Y. Zhang et al., “Brain tumor segmentation using multi-cascaded convolutional neural networks and conditional random field,” IEEE Access, vol. 7, pp. 92615–92629, 2019.
View at: Publisher Site | Google Scholar
M. Kistler, S. Bonaretti, M. Pfahrer, R. Niklaus, and P. Büchler, “The virtual skeleton database: an open access repository for biomedical research and collaboration,” Journal of Medical Internet Research, vol. 15, no. 11, p. e245, 2013.
View at: Publisher Site | Google Scholar
S. Khan, N. Islam, Z. Jan, I. Ud Din, and J. J. P. C. Rodrigues, “A novel deep learning based framework for the detection and classification of breast cancer using transfer learning,” Pattern Recognition Letters, vol. 125, pp. 1–6, 2019.
View at: Publisher Site | Google Scholar
P. Kaur, G. Singh, and P. Kaur, “Intellectual detection and validation of automated mammogram breast cancer images by multi-class SVM using deep learning classification,” Informatics in Medicine Unlocked, vol. 16, p. 100151, 2019.
View at: Publisher Site | Google Scholar
Z. Wang, M. Li, H. Wang et al., “Breast cancer detection using extreme learning machine based on feature fusion with CNN deep features,” IEEE Access, vol. 7, pp. 105146–105158, 2019.
View at: Publisher Site | Google Scholar
Y. Yang, J. Liu, and X. Zhou, “A CRISPR-based and post-amplification coupled SARS-CoV-2 detection with a portable evanescent wave biosensor,” Biosensors and Bioelectronics, vol. 190, Article ID 113418, 2021.
View at: Publisher Site | Google Scholar
J. Manokaran, F. Zabihollahy, A. Hamilton-Wright, and E. Ukwatta, “Deep learning-based detection of COVID-19 from chest X-ray images,” in Proceedings of the Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging, Houston, TX, USA, February 2021.
View at: Publisher Site | Google Scholar
E. Luz, P. L. Silva, R. Silva et al., “Towards an effective and efficient deep learning model for covid-19 patterns detection in X-ray images,” Research on Biomedical Engineering, vol. 7, pp. 1–14, 2021.
View at: Google Scholar
B. Ghoshal and A. Tucker, “Estimating uncertainty and interpretability in deep learning for coronavirus (COVID-19) detection,” 2020, http://arxiv.org/abs/2003.10769.
View at: Google Scholar
A. Narin, C. Kaya, and Z. Pamuk, “Automatic detection of coronavirus disease (COVID-19) using X-ray images and deep convolutional neural networks,” 2020, http://arxiv.org/abs/2003.10849.
View at: Google Scholar
J. Zhang, Y. Xie, Y. Li, C. Shen, and Y. Xia, “COVID-19 screening on chest X-ray images using deep learning based anomaly detection,” 2020, http://arxiv.org/abs/2003.12338.
View at: Google Scholar
M. Zeraati, R. Pourmohamad, B. Baghchi, N. P. Singh Chauhan, and G. Sargazi, “Optimization and predictive modelling for the diameter of nylon-6,6 nanofibers via electrospinning for coronavirus face masks,” Journal of Saudi Chemical Society, vol. 25, no. 11, Article ID 101348, 2021.
View at: Publisher Site | Google Scholar
X. Li, Z.-Q. Dong, P. Yu et al., “Effect of self-assembly on fluorescence in magnetic multiphase flows and its application on the novel detection for COVID-19,” Physics of Fluids, vol. 33, no. 4, p. 042004, 2021.
View at: Publisher Site | Google Scholar
L. Huang, R. Han, T. Ai et al., “Serial quantitative chest CT assessment of COVID-19: a deep learning approach,” Radiology: Cardiothoracic Imaging, vol. 2, no. 2, Article ID e200075, 2020.
View at: Publisher Site | Google Scholar
Y. Oh, S. Park, and J. C. Ye, “Deep learning COVID-19 features on CXR using limited training data sets,” IEEE Transactions on Medical Imaging, vol. 39, no. 8, pp. 2688–2700, 2020.
View at: Publisher Site | Google Scholar
W. Liang, J. Yao, A. Chen et al., “Early triage of critically ill COVID-19 patients using deep learning,” Nature Communications, vol. 11, no. 1, 2020.
View at: Publisher Site | Google Scholar
M. Abadi, Q. Hasan, S. Rahmati, A. Sharifi, and M. Ahmadi, “HSSAGA: designation and scheduling of nurses for taking care of COVID-19 patients using novel method of hybrid salp swarm algorithm and genetic Algorithm,” Applied Soft Computing, vol. 108, Article ID 107449, 2021.
View at: Publisher Site | Google Scholar
M. Ahmadi, A. Sharifi, S. Dorosti, S. Jafarzadeh Ghoushchi, and N. Ghanbari, “Investigation of effective climatology parameters on COVID-19 outbreak in Iran,” Science of the Total Environment, vol. 729, Article ID 138705, 2020.
View at: Publisher Site | Google Scholar
Nasirpour, M. Hossein, A. Sharifi, M. Ahmadi, and S. Jafarzadeh Ghoushchi, “Revealing the relationship between solar activity and COVID-19 and forecasting of possible future viruses using multi-step autoregression (MSAR),” Environmental Science and Pollution Research, vol. 28, pp. 1–11, 2021.
View at: Publisher Site | Google Scholar
A. Davoudi, M. Ahmadi, A. Sharifi et al., “Studying the effect of taking statins before infection in the severity reduction of COVID-19 with machine learning,” BioMed Research International, vol. 2021, Article ID 9995073, 12 pages, 2021.
View at: Publisher Site | Google Scholar
A. Sharifi, M. Ahmadi, and A. Ali, “The impact of artificial intelligence and digital style on industry and energy post-COVID-19 pandemic,” Environmental Science and Pollution Research, vol. 28, pp. 1–21, 2021.
View at: Publisher Site | Google Scholar
A. Memiş, S. Varlı, and F. Bilgili, “Semantic segmentation of the multiform proximal femur and femoral head bones with the deep convolutional neural networks in low quality MRI sections acquired in different MRI protocols,” Computerized Medical Imaging and Graphics: The Official Journal of the Computerized Medical Imaging Society, vol. 81, Article ID 101715, 2020.
View at: Publisher Site | Google Scholar
A. Duran, P. M. Jodoin, and C. Lartizien, “Prostate cancer semantic segmentation by gleason score group in bi-parametric MRI with self attention model on the peripheral zone,” in Proceedings of the Medical Imaging with Deep Learning, pp. 193–204, PMLR, Montreal, Canada, September 2020.
View at: Google Scholar
N. Alalwan, A. Abozeid, A. A. ElHabshy, and A. Alzahrani, “Efficient 3d deep learning model for medical image semantic segmentation,” Alexandria Engineering Journal, vol. 60, no. 1, pp. 1231–1239, 2021.
View at: Publisher Site | Google Scholar
G. Karayegen and M. F. Aksahin, “Brain tumor prediction on MR images with semantic segmentation by using deep learning network and 3D imaging of tumor region,” Biomedical Signal Processing and Control, vol. 66, Article ID 102458, 2021.
View at: Publisher Site | Google Scholar
M. Ahmadi, A. Sharifi, S. Hassantabar, and S. Enayati, “QAIS-DSNN: tumor area segmentation of MRI image with optimized quantum matched-filter technique and deep spiking neural network,” BioMed Research International, vol. 2021, Article ID 6653879, 16 pages, 2021.
View at: Publisher Site | Google Scholar
M. Ahmadi, A. Sharifi, M. Jafarian Fard, and N. Soleimani, “Detection of brain lesion location in MRI images using convolutional neural network and robust PCA,” International Journal of Neuroscience, pp. 1–12, 2021.
View at: Publisher Site | Google Scholar
D. Tabernik, M. Kristan, and A. Leonardis, “Spatially-adaptive filter units for compact and efficient deep neural networks,” International Journal of Computer Vision, vol. 128, no. 8, pp. 2049–2067, 2020.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2021 Amin Valizadeh and Morteza Shariatee. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

908

Downloads

1041

Citations