Skip to main content
. 2020 Aug 28;7:283. doi: 10.1038/s41597-020-00622-y

Table 1.

An overview of existing GI datasets.

Dataset Findings Size Availability
CVC-35618 Polyps 356 images by request
CVC-ClinicDB19 (also named CVC-612) Polyps 612 images open academic
CVC-VideoClinicDB18 (also named CVC-12k) Polyps 11954 images by request
CVC-ColonDB62 Polyps 380 imagesψ by request
Endoscopy Artifact detection 201963 Endoscopic Artifacts 5,138 images open academic
ASU-Mayo polyp database20 Polyps 18,781 images by request
ETIS-Larib Polyp DB64 Polyps 196 images open academic
KID65 Angiectasia, bleeding, inflammations, polyps 2371 images and 47 videos open academic
GIANA 201766 Polyps & Angiodysplasia 3462 images and 38 videos by request
GIANA 201867,68 Polyps & Small bowel lesions 8262 images and 38 videos by request
GASTROLAB69 GI lesions Some 100s of images and few videos open academic
WEO Clinical Endoscopy Atlas70 GI lesions 152 images by request
GI Lesions in Regular Colonoscopy Data Set71 GI lesions 76 images by request
Atlas of Gastrointestinal Endoscope72 GI lesions 1295 images unknown

El salvador atlas of gastrointestinal

video endoscopy73

GI lesions 5071 video clips open academic
Kvasir22 Polyps, esophagitis, ulcerative colitis, Z-line, pylorus, cecum, dyed polyp, dyed resection margins, stool 8000 images open academic
Kvasir-SEG49 Polyps 1000 images open academic
Nerthus74 Stool - categorization of bowel cleanliness 21 videos open academic

Including ground truth segmentation masks. Video capsule endoscopy. Not available anymore. ψContour.

Not really a dataset usable for machine learning. It is more a medical atlas or database for education with a several low-quality samples of various findings in the GI tract.