SA-UNet: Spatial Attention U-Net for Retinal Vessel Segmentation

The precise segmentation of retinal blood vessels is of great significance for early diagnosis of eye-related diseases such as diabetes and hypertension. In this work, we propose a lightweight network named Spatial Attention U-Net (SA-UNet) that does not require thousands of annotated training samples and can be utilized in a data augmentation manner to use the available annotated samples more efficiently... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK BENCHMARK
Retinal Vessel Segmentation CHASE_DB1 SA-UNet F1 score 0.8153 # 1
AUC 0.9905 # 1
Retinal Vessel Segmentation DRIVE SA-UNet F1 score 0.8263 # 1
AUC 0.9864 # 1
Accuracy 0.9698 # 1

Methods used in the Paper