|
| Reference | Year | Method | Advantages | Disadvantages |
|
| Bourdev et al. [28] | 2010 | HOG | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be used in a wide range of applications |
| Xia et al. [27] | 2005 | HOG | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be implied in an extended variety of applications |
| Lowe [29] | 2004 | SIFT | Accurate identification of areas and edges | High computational complexity, inability to be implied in an extended variety of applications |
| He and Wang [30] | 1990 | LBP | High capability in cryptography and image data encryption and edge detection operations | The algorithm is slow |
| Bay et al. [31] | 2008 | SURF | Accurate identification of areas and edges | High computational complexity, inability to be used in a wide range of applications |
| Derpanis [32] | 2004 | Harris corner detection | Ability to find the corners of an image outside the edges, ability to be used in a wide range of high-sensitivity image processing systems | Low accuracy and slow method |
| Shi and Tomasi [33] | 1994 | Shi-Tomasi | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be used in a wide range of applications |
| Medioni and Yasumoto [34] | 1987 | Corner detection with subpixels | Ability to find the corners of an image outside the edges, ability to be used in a wide range of high-sensitivity image processing systems | Low accuracy and slow method |
| Smith and Brady [35] | 1997 | SUSAN corner detection | Accurate edge detection based on texture and brightness and better capabilities than classic operators such as Canny and Prewitt | Low accuracy in high-resolution images and slow method |
| Rosten and Drummond [36] | 2005 | FAST | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be used in a wide range of applications |
| Rosten et al. [37] | 2010 | FAST-ER | Ability to specify areas, especially at the edges, with high resolution and precision in multiscale modes | High computational complexity, inability to be used in a wide range of applications |
| Mair et al. [38] | 2010 | AGAST | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be used in a wide range of applications |
| Leutenegger et al. [39] | 2011 | Multiscale AGAST | Ability to specify areas, especially at the edges, with high resolution and precision in multiscale modes | High computational complexity, inability to be used in a wide range of applications |
| Venegas-Barrera and Manjarrez [40] | 2004 | BOW | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be used in a wide range of applications |
| Brox et al. [41] | 2011 | Poselets | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be used in a wide range of applications |
| Zhu et al. [42] | 2005 | Textons | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be used in a wide range of applications |
| Strauss and Hartigan [47] | 1975 | K-means | Ability to find clusters in images and cluster them | Slow method and need to combine with faster methods |
| Shen et al. [54] | 2004 | Edge detection and regional growth | Ability to specify areas, especially at the edges, with high clarity and accuracy | High computational complexity, inability to be used in a wide range of applications |
| Shan et al. [55] | 2004 | SVM | High capability in high-precision image classification operations in pairs and the ability to separate features with vectors | High computational complexity, slow method |
| Shotton et al. [56] | 2006 | MRF | High capability in high-precision image classification operations in pairs and the ability to separate features with vectors | High computational complexity, slow method |
| Hassantabar et al. [61] | 2020 | CNN | Ability to diagnose the COVID-19 infected lung tissue for segmentation and classification of patients | (i) Small numbers of images
(ii) Unable to find illness severity |
| Dorosti et al. [62] | 2020 | Sensitivity analysis | This approach can help in the identification of beneficial parameters as well as the avoidance of patient mortality in all sorts of disease | (i) Data limit
(ii) Ignoring other variables |
| Sharifi et al. [63] | 2021 | CNN | Diagnosis of fatigue foot using CNN | (i) High computational complexity
(ii) Integrated only on CNN method |
| Laradji et al. [64] | 2021 | Supervised consistency learning | The best loss function for prediction | (i) Unable to detect patients uniquely
(ii) Should be connected to other methods |
|
