Table 7.

Analysis of BC diagnosis based on DL and NN approaches in mammography using information from reviewed articles.

References	Approach	Targets for ML	Targets for BC	Advantages of the ML approach	Advantages of the BC approach	Disadvantages	Evaluation criteria
							AUC	ACC	SPE	SEN	Recall	F-Score
[100]	CNN	•Proposing an approach based on NNs to recognize images with high accuracy	•Presenting a model of classification and accuracy that is close to that of a human performance	•Detects and classifies benign or malignant lesions on mammograms without human intervention	•Exceptional success in the classification of images •Assisting radiologists in mammography images analysis	•Inapplicability of the presented approach to large datasets	✔	✘	✘	✘	✘	✘
[101]	CNN	•Proposed a system to improve the diagnostic performance of breast lesions using an integrated detection and classification method	•Finding a mechanism to improve classification accuracy	•Increasing accuracy in the process of identifying BC lesions	•Assisting professionals in accurately diagnosing breast lesions •Achieve promising breast lesion diagnostic performance.	•A lack of effective clinical parameters for diagnosis	✘	✔	✘	✘	✘	✔
[102]	CNN	•Proposing a neural network-based method for extracting high-level semantic features	•Optimizing classification performance through the use of selected features	•The proposed approach is superior to other CNN models and obtains presenting better performance in classification	•Improvements in the classification of images •Enhancements to the extraction process of high-level semantic features	•Limiting the extraction of features only from mammograms •Inability to use newly developed imaging features as well as more clinical characteristics to diagnose BC grades.	✔	✔	✘	✘	✘	✘
[103]	DAL	•Proposed a novel mass detection methodology with DAL	•Improve the accuracy of the classification process	•Minimizing the annotation efforts in detecting of BC •Compared to its counterparts, the proposed learning framework is more effective at detecting BC	•Breast mass accurate detection in mammogram •Reduces the requirement of the annotated samples in detection.	•Failure to perform more measurements in sample selection criteria. •Failure to use the proposed approach in various imaging types, such as pathology and MRI	✘	✘	✘	✔	✘	✘