Densely connected U‐Net retinal vessel segmentation algorithm based on multi‐scale feature convolution extraction

Purpose The segmentation results of retinal blood vessels have a significant impact on the automatic diagnosis various ophthalmic diseases. In order to further improve accuracy vessels, we propose an improved algorithm based multiscale vessel detection, which extracts features through densely connected networks and reuses features. Methods A parallel fusion serial embedding feature dense connection U-Net structure are designed. method, input images extracted for Inception convolution block convolution, respectively, then fused into subsequent network. mode, is embedded in network module, used replace classical encoder part, so as achieve extraction efficient utilization complex thereby performance. Results experimental analysis standard DRIVE CHASE_DB1 databases shows that sensitivity, specificity, accuracy, AUC methods reach 0.7854, 0.9813, 0.9563, 0.9794; 0.7876, 0.9811, 0.9565, 0.9793 0.8110, 0.9737, 0.9547, 0.9667; 0.8113, 0.9717, 0.9574, 0.9750, respectively. Conclusions show detection can effectively enhance model's ability detect performance, superior some mainstream algorithms at present.