Applications of AI/ML for medical imaging.
| Area | Modalities | Classification | Detection | Diagnosis | Extraction | Grading | Identification | Prediction | Segmentation | AI/ML Families |
| Brain | MRI, fMRI, PET, and DTI | X | X | X | DBN, SBM, CNN, and FCN | |||||
| Retinal | Fundus | X | X | DNN, FCN, cGAN, and SSD | ||||||
| Chest | X-ray, CT, and MRI | X | X | X | CNN, RCNN, U-NET, and DCNN | |||||
| Digital pathology | H&E WSI | X | X | X | X | U-NET, TCGA, and GRAG | ||||
| Breast | Mammogram, MRI, US, and DBT | X | X | CNN, cGAN, and U-NET | ||||||
| Cardiac | MRI | X | CNN, DNN, U-NET, and cGAN | |||||||
| Abdominal | CT, MRI, and endoscopic | X | X | U-NET, CNN, and DNN | ||||||
| Dermatology | Dermascope and photograph | X | X | X | SLSNet and CNN |
Abbreviations: AI, artificial intelligence; cGAN, conditional generative adversarial network; CNN, convolutional neural networks; CT, computed tomography; DBN, deep belief networks; DBT, deep breast tomosynthesis; DCNN, deep CNN; DNN, deep neural network; DTI, diffusion tensor imaging; FCN, fully CNN; fMRI, functional MRI; H&E, hematoxylin and eosin; ML, machine learning; MRI, magnetic resonance imaging; PET, positron electron microscopy; RCNN, region-based CNN; SSD, single-shot multibox detector; TCGA, The Cancer Genome Atlas; US, ultrasonography; WSI, whole slide image.