Self-relabeling for noise-tolerant retina vessel segmentation through label reliability estimation

• Published in: BMC medical imaging Vol. 22; no. 1; p. 8
• Main Authors: Li, Jiacheng, Li, Ruirui, Han, Ruize, Wang, Song
• Format: Journal Article
• Language: English
• Published: England BioMed Central 12.01.2022; BMC
• Subjects: Accuracy; Annotations; Background noise; Blood vessels; Classification; Datasets; Deep Learning; Errors; Humans; Image processing; Image Processing, Computer-Assisted - methods; Image segmentation; Label map correction; Labeling; Labels; Medical research; Methods; Noise; Noise-tolerant; Pixels; Reliability estimation; Retina; Retina image segmentation; Retinal images; Retinal Vessels - diagnostic imaging; Temporal statistics; Training; Reliability estimation; Noise-tolerant; Label map correction; Temporal statistics; Retina image segmentation

Retinal vessel segmentation benefits significantly from deep learning. Its performance relies on sufficient training images with accurate ground-truth segmentation, which are usually manually annotated in the form of binary pixel-wise label maps. Manually annotated ground-truth label maps, more or l...