Figure 2 | Scientific Reports

Figure 2

From: Density clustering-based automatic anatomical section recognition in colonoscopy video using deep learning

Figure 2

Training data acquisition process: Set (A): The number of images selected after frame sampling. Each anatomical segment is clipped using the timestamp as a reference. Set (B): The number of informative images listed according to the corresponding anatomical segment. After the candidate data is reviewed by clinicians, 2547 AO images, 2773 flexure images, and 15,520 outside-of-the-body images were obtained. (AO: appendiceal orifice, ICV: ileocecal valve, HF: hepatic flexure, SF: splenic flexure, SDJ: descending colon–sigmoid colon junction, RSJ: rectum–sigmoid colon junction, A-colon: ascending colon, T-colon: transverse colon, D-colon: descending colon).

Back to article page