CGDD-Net: Context-Guided Dynamic Detail Modeling for Retinal Vessel Segmentation

2026-10-04

AI Summary

This study addresses the inherent trade-off between geometric feature capture and fine-grained reconstruction preservation in retinal vessel segmentation by proposing CGDD-Net, a lightweight network. Methodologically, it incorporates a CSDE deformable encoder, SAMG multi-kernel gating, and DCDF dynamic cross-scale fusion modules to enable adaptive feature extraction. Furthermore, it introduces a shared intermediate detail fusion mechanism that integrates multi-scale features prior to decoding as an alternative to conventional skip connections, alongside a context-guided dynamic detail modeling connection that shares the decoding path. Experimental evaluations demonstrate that with only 1.96 million parameters, the proposed model achieves a maximum AUC of 0.9938 across four public datasets, significantly outperforming existing baseline models while maintaining superior parameter efficiency.

Abstract

Accurate retinal vessel segmentation requires features that capture vascular geometry while preserving information for fine-scale reconstruction. We propose CGDD-Net, a context-guided dynamic detail modeling network that connects adaptive feature extraction to a shared decoder pathway. Context-Guided Scale-Adaptive Deformable Encoding (CSDE) combines fixed-grid convolution with deformable local attention to capture vascular patterns at multiple spatial extents. Spatially Adaptive Multi-Kernel Gating (SAMG) selects receptive-field responses at each location. Dynamic Cross-Scale Detail Fusion (DCDF) aligns the gated intermediate features and compresses them into an eight-channel representation, which is reused at three decoder resolutions together with selected encoder skips. This design consolidates intermediate information before decoding instead of transferring each middle-stage feature through a separate direct skip. On DRIVE, CHASE\_DB1, STARE, and HRF, CGDD-Net achieves AUC values of 0.9824, 0.9938, 0.9895, and 0.9874, with F1 scores of 0.8323, 0.8102, 0.8510, and 0.8157, respectively. The complete model contains 1.96 million trainable parameters. In cumulative ablations, the full model improves F1 over the internal baseline by 2.48, 0.61, and 3.49 percentage points on DRIVE, CHASE\_DB1, and STARE. Twelve directed cross-dataset experiments further characterize transfer without target-domain adaptation. The results support shared intermediate detail delivery as an effective, parameter-compact architecture for retinal vessel segmentation. Code is available at \url{https://github.com/lixincheng-xcl/CGDD-Net}.

Problem

• Retinal Vessel Segmentation
• Vascular Geometry
• Fine-scale Reconstruction
• Feature Extraction

Innovation

• Context-Guided Deformable Encoding
• Multi-Kernel Gating
• Cross-Scale Detail Fusion
• Retinal Vessel Segmentation
• Parameter-Efficient Architecture