| Color ocular fundus images | 6679 random sampling images from Kaggle’s Diabetic Retinopathy Detection | Detect retinal hemorrhages | CNN | AUC of 0.894 and 0.972 | Ophthalmology | [95] |
Ocular fundus images | 168 images with glaucoma and 428 control | System to detect and evaluate glaucoma | CNN: ResNet and U-Net | AUC of 0.91 and 0.84 respectively | Ophthalmology | [98] | |
Ocular fundus images | 90,000 images with their diagnoses | Predict the evolution of diabetic retinopathy with fundus images | CNN | AUC of 0.95 | Ophthalmology | [155] | |
Fundus images | 7000 colour fundus images | Image quality in the context of diabetic retinopathy | CNN | Accuracy of 100% | Ophthalmology | [156] | |
AREDS (age related eye disease study) image | 130,000 fundus images | Diagnosis of Age-related Macular Degeneration | CNN | 94.97 sensitivity and 98.32% specificity | Ophthalmology | [157] | |
Fundus images | 219,302 from normal participants without hypertension, diabetes mellitus (DM), and any smoking history | Predict age and sex from retinal fundus images | CNN | AUC 0.96 | Ophthalmology | [2] ; [158] | |
Dermoscopy | Dermoscopy images | 350 images of melanomas and 374 benign nevus | Dermoscopy CAD system for acral lentiginous melanoma diagnosis | CNN | Accuracy of over 0.80 | Oncology | [2] ; [99] |
Patient demographics and clinical images | 49,567 images | Recognize nails nychomycosis lesions | Region-based- CNN | AUC of 0.98, AUC of 0.95, AUC of 0.93, AUC value of 0.82 in the different datasets | Dermatology | [120] | |
Stress 99mTc-sestamibi or tetrofosmin myocardial perfusion images | 1638 patients | Obstructive coronary disease automatic prediction system | CNN | Sensitivity value of 0.82 and 0.69 for both use cases | Cardiovascular | [159] | |
Arterial labeling | Arterial spin labeling (ASL) perfusion images | 140 subjects | Monitoring cerebral arterial perfusion via spin labeling | CNN | AUC of 0.94 | Cardiovascular | [91] |
Frames from endoscopy | Frames from endoscopy videos | 205 normal and 360 abnormal images | Detection and localization system of gastrointestinal anomalies via endoscopy | CNN | AUC of over 0.80 | Gastroenterology | [103] |