Radiography | Radiography images | 7821 subjects with 6 monitoring phases | CAD for diagnosis of knee osteoarthritis | Deep Siamese | Accuracy of 0.66 | Traumatology | [115] |
Radiography images | 420 radiography images (219 control group, 201 ostearthritis) [2] | Radiographies CAD for hip osteoarthritis diagnosis | CNN | Accuracy of 0.92 | Traumatology | [153] | |
Radiography images | 112,120 frontal view chest radiographs from 30,805 patients and 17,202 frontal view chest radiographs with a binary class label for normal vs abnormal | Abnormality detection in chest radiographs | CNN | AUROCs of 0.960 and 0.951. AUROCs of 0.900 and 0.893 | Radiology | [145] | |
Slide image | Pathology cancer images (hematoxylin and eosin) | 5202 images tumor infiltrating lymphocytes | Study of tumor tissue samples. Localize areas of necrosis and lymphocyte infiltration | Two CNNs | AUC of 0.95 | Oncology | [2] ; [76] |
Giemsa-stained thin blood smear slides cell images | 27,558 cell images 150 infected and 50 healthy patients | Create a screening system for Malaria | CNN | Accuracy of 0.94 | Infectious Disease | [80] | |
Microscopy image patches | 249 images belonging to 20 histologic categories | Classification of breast cancer histology microscopy images | CNN with a Support Vector Machine (SVM) | Accuracy of 0.77 for four class classification and an accuracy of 0.83 for carcinoma /noncarcinoma classification | Oncology | [109] | |
Microscopy histopathological images | 7909 images of eight subclasses of breast cancers | CAD for breast cancer histopathological diagnosis | CNN | Accuracy of 0.93 | Oncology | [110] | |
Microscope images | 200 female subjects aged from 22 to 64 | Cervix cancer screening | Multiscale CNN | Mean and standard deviation of 0.95 and 0.18 | Oncology | [122] | |
Whole-slide prostate histopathology images | 2663 images from 32 whole slide prostate histopathology images | Whole-slide histopathology images to outline the malignant regions | CNN | Dice coefficient of 0.72 | Oncology | [2] ; [154] | |
Ocular fundus | 2D Ocular fundus images | 243 retina images | Diagnose retinal lesions | CNN | Precision recall curve of 0.86 in microaneurysms and 0.64 in exudates | Ophthalmology | [2] ; [78] |
Ocular fundus images 2D | Over 85,000 images | Diabetic retinopathy detection and stage classification | Bayesian CNN | AUC value of 0.99 | Ophthalmology | [88] |