Author Methods C S ADS Dataset Size Type Deepak and Ameer [2 ] GoogLeNet with transfer learning √ — — 3064 brain MRI images from 233 patients (diagnosed meningioma, glioma, and pituitary tumors) are with T1-CE MRI modality and include coronal, sagittal, and axial views 512 × 512 .Mat file Hu et al. [3 ] Multikernel depthwise convolution √ — — Chest X-ray 14 dataset and Chest X-ray 2017 dataset 512 × 512 preprocess to 256 × 256 Dicom Liu et al. [4 ] Landmark-based deep multi-instance learning (LDMIL) √ — — 1.5T and 3T T1-weighted structural MRI data (1526) Voxel level with 256 × 256 × 256 — Suk et al. [5 ] Deep ensemble learning of sparse regression models √ — — 805 subjects of 186 (AD), 393 (MCI), and 226 (normal control, NC) Voxel level with 256 × 256 × 256 Neuroimaging Informatics Technology Initiative (NIfTI) Xue et al. [6 ] AlexNet, VGG-19, and ResNet √ — — 420 anteroposterior (AP) view hip X-ray images with normal (219) and OA (201) 432 mm × 356 mm, with a pixel resolution of 0.187 mm × 0.187 mm — Safdar et al. [8 ] YOLO √ — — T1-weighted and FLAIR images from MRI images of patients who suffered from low-grade glioma 256 × 256 Dicom Ronneberger et al. [9 ] UNet — √ — Transmission electron microscopy of the Drosophila first instar larva ventral nerve cord (VNC) 512 × 512 — Weng et al. [11 ] UNet — √ — MRI, CT, and ultrasound (Promise12, chaos, NERVE dataset) 256 × 256; 512 × 512; 580 × 420 Dicom Li et al. [12 ] UNet — √ — NIH pancreatic segmentation dataset that contains 82 contrast-enhanced abdominal CT volumes and corresponding fine annotations 512 × 512 × L, where L ∈ [181, 466] — Pravitasari et al. [10 ] UNet-VGG16 — √ — Real dataset from the general hospital (RSUD) Dr. Soetomo (152) 256 × 256 Dicom Bisa [14 ] R-CNN √ √ √ BraTS (brain tumor segmentation) challenge dataset — Neuroimaging Informatics Technology Initiative (NIfTI) Süleyman Yıldırım and Dandıl [15 ] Mask R-CNN √ √ √ MR image of multiple sclerosis dataset (1838) min 256, max 512 — He et al. [16 ] Mask R-CNN √ √ √ COCO dataset 640 × 800 JPG