Papers › Dual encoding feature filtering generalized attention UNET for retinal vessel segmentation

Dual encoding feature filtering generalized attention UNET for retinal vessel segmentation

2 Jun 2025arXiv:2506.02312archive 2025-07-28

Md Tauhidul Islam, Wu Da-Wen, Tang Qing-Qing, Zhao Kai-Yang, Yin Teng, Li Yan-Fei, Shang Wen-Yi, Liu Jing-Yu, Zhang Hai-Xian

Retinal blood vessel segmentation is crucial for diagnosing ocular and cardiovascular diseases. Although the introduction of U-Net in 2015 by Olaf Ronneberger significantly advanced this field, yet issues like limited training data, imbalance data distribution, and inadequate feature extraction persist, hindering both the segmentation performance and optimal model generalization. Addressing these critical issues, the DEFFA-Unet is proposed featuring an additional encoder to process domain-invariant pre-processed inputs, thereby improving both richer feature encoding and enhanced model generalization. A feature filtering fusion module is developed to ensure the precise feature filtering and robust hybrid feature fusion. In response to the task-specific need for higher precision where false positives are very costly, traditional skip connections are replaced with the attention-guided feature reconstructing fusion module. Additionally, innovative data augmentation and balancing methods are proposed to counter data scarcity and distribution imbalance, further boosting the robustness and generalization of the model. With a comprehensive suite of evaluation metrics, extensive validations on four benchmark datasets (DRIVE, CHASEDB1, STARE, and HRF) and an SLO dataset (IOSTAR), demonstrate the proposed method's superiority over both baseline and state-of-the-art models. Particularly the proposed method significantly outperforms the compared methods in cross-validation model generalization.

PaperPDFCode

Code

TauhidScu/DEFFA-Unet officialpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Data AugmentationRetinal Vessel Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Retinal Vessel Segmentation CHASE_DB1 DEFFA-Unet 1:1 Accuracy 0.9712 #9 of 16 Archive leaderboard report
Retinal Vessel Segmentation CHASE_DB1 DEFFA-Unet AUC 0.9823 #9 of 16 Archive leaderboard report
Retinal Vessel Segmentation CHASE_DB1 DEFFA-Unet Average IOU 0.6891 #9 of 16 Archive leaderboard report
Retinal Vessel Segmentation CHASE_DB1 DEFFA-Unet DSC 0.8156 #9 of 16 Archive leaderboard report
Retinal Vessel Segmentation CHASE_DB1 DEFFA-Unet MCC 0.7892 #9 of 16 Archive leaderboard report
Retinal Vessel Segmentation DRIVE DEFFA-Unet 1:1 Accuracy 0.9701 #5 of 22 Archive leaderboard report
Retinal Vessel Segmentation DRIVE DEFFA-Unet AUC 0.9861 #5 of 22 Archive leaderboard report
Retinal Vessel Segmentation DRIVE DEFFA-Unet Average IOU 0.7154 #5 of 22 Archive leaderboard report
Retinal Vessel Segmentation DRIVE DEFFA-Unet DSC 0.8347 #5 of 22 Archive leaderboard report
Retinal Vessel Segmentation DRIVE DEFFA-Unet MCC 0.8079 #5 of 22 Archive leaderboard report
Retinal Vessel Segmentation HRF DEFFA-Unet 1:1 Accuracy 0.9723 #2 of 4 Archive leaderboard report
Retinal Vessel Segmentation HRF DEFFA-Unet AUC 0.9845 #2 of 4 Archive leaderboard report
Retinal Vessel Segmentation HRF DEFFA-Unet Average IOU 0.7089 #2 of 4 Archive leaderboard report
Retinal Vessel Segmentation HRF DEFFA-Unet DSC 0.8289 #2 of 4 Archive leaderboard report
Retinal Vessel Segmentation HRF DEFFA-Unet MCC 0.8012 #2 of 4 Archive leaderboard report
Retinal Vessel Segmentation STARE DEFFA-Unet 1:1 Accuracy 0.9689 #6 of 10 Archive leaderboard report
Retinal Vessel Segmentation STARE DEFFA-Unet AUC 0.9834 #6 of 10 Archive leaderboard report
Retinal Vessel Segmentation STARE DEFFA-Unet Average IOU 0.7012 #6 of 10 Archive leaderboard report
Retinal Vessel Segmentation STARE DEFFA-Unet DSC 0.8234 #6 of 10 Archive leaderboard report
Retinal Vessel Segmentation STARE DEFFA-Unet MCC 0.7945 #6 of 10 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Concatenated Skip ConnectionConvolutionMax PoolingReLUU-Net

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections