An Approach for Thoracic Syndrome Classification with Convolutional Neural Networks

Table 1

Brief summary of inspiration from the earlier research in the domain of disease identification.

S. No.	Author and year	Paper title	Technique used	Objective
1	Bharati & Podder; [30]	Disease Detection from Lung X-ray Images Based on Hybrid Deep Learning Subrato	CNN, vanilla NN	The model proposed classification of chest diseases with its metrics as precision, recall, and -score
2	Rajaraman et al. [31]	Assessment of an Ensemble of Machine Learning Models towards Abnormality Detection in Chest Radiographs	Sequential CNN	Used weighted averaging to in base learners to classify the chest - rays
3	Chan et al. [28]	Effective Pneumothorax Detection for Chest X-Ray Images Using Local Binary Pattern and Support Vector Machine	Support vector machine and local binary pattern	The paper proposed a methodology to detect the lung diseases using the local binary patterns and then further used the SVM technique to classify the type of disease
4	Li et al. [32]	Thoracic Disease Identification and Localization with Limited Supervision	CNN	Identification and localization of abnormalities in the X-rays
5	Sharma et al. [33]	An Analysis Of Convolutional Neural Networks For Image Classification	CNN	The paper focusses on the analysis of real time images of three types of CNN’s; these are AlexNets, GoogLeNet, and ResNet50
6	Yao et al. [34]	Learning to diagnose from scratch by exploiting dependencies among labels	LSTM	Used long short-term memory networks for distinction between chest diseases
7	Esteva et al. [35]	Dermatologist-Level Classification of Skin Cancer with Deep Neural Networks	t-SNE-based NN	Analyzed the internal features of the cells by using the CNN with the t-distributed stochastic neighbor embedding
8	Wang et al. [27]	ChestX-ray8: Hospital-Scale Chest X-Ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases	CNN	The work focusses on how thoracic ailments can be discovered and explicitly located with the help of a combined softly supervised multilabelled image sorting and ailment localization framework; the same is verified with the dataset used in the paper