EDCE-Net: Edge Detection and Connectivity Enhancement Network for Retinal Vessel Segmentation
Bibliographic record
Abstract
Utilizing deep learning methods for retinal vessel segmentation is crucial for aiding ophthalmologists in diagnosing fundus diseases clinically.However, existing deep learning models often fail to achieve ideal performance when facing complex scenarios such as multidirectional vascular endings and overlapping intersections in the retina.In order to improve the detection rate of retinal blood vessels and ensure high connectivity, this paper proposes a new method EDCE-Net, which combines the directional information and scale differences of retinal blood vessels for the first time.Specifically, by designing an edge detection module (EDM) to assist the model in extracting vascular details and directional information at different scales, and integrating them into the depth feature map, it ensures the restoration of lost edge information while maintaining the correctness of the global topology.Additionally, this paper designs a connectivity enhancement module (CEM) based on multiscale coordinate attention to effectively assist the model in establishing long-range spatial dependencies in spatial directions, accurately capturing and locating crossing and overlapping vessel structures, so as to improve the connectivity of vessel segmentation.Finally, we employ multiple auxiliary loss functions to provide hierarchical supervision for the model, fully considering the impact of various levels of feature maps on the segmentation outcomes, thereby enhancing the robustness of the model.EDCE-Net was validated on four public datasets: DRIVE, CHASE_DB1, STARE, and DCA1, achieving AUC values of 98.79%, 98.76%, 98.06%, and 99.08%, respectively, which are the highest compared to several existing advanced methods.JI reached 70.15%, 68.84%, 70.52%, and 68.37% on these datasets, respectively, representing improvements of 2.51%, 3.58%, 2.06%, and 5.3% compared to the baseline, highlighting the potential of the proposed method in aiding retinal disease diagnosis.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".