RetVes segmentation: A pseudo-labeling and feature knowledge distillation optimization technique for retinal vessel channel enhancement

Favour Ekong, Yongbin Yu, Rutherford Agbeshi Patamia, Kwabena Sarpong, Chiagoziem Chima Ukwuoma, Akpanika Robert Ukot, Jingye Cai

Abstract: Highlights•We propose a framework for retinal vessel segmentation model with pseudo-labelling.•Comprehensive pipeline: Includes enhancement, segmentation, and domain adaptation.•Outstanding performance: Accurate vessel segmentation with minimal errors.•Unique contrastive pretraining submodule for improved vessel feature learning.•Robustness: Improved model generalisation, addressing missing annotations.