Global Transformer and Dual Local Attention Network via Deep-Shallow Hierarchical Feature Fusion for Retinal Vessel Segmentation

• Published in: IEEE transactions on cybernetics Vol. 53; no. 9; pp. 5826 - 5839
• Main Authors: Li, Yang, Zhang, Yue, Liu, Jing-Yu, Wang, Kang, Zhang, Kai, Zhang, Gen-Sheng, Liao, Xiao-Feng, Yang, Guang
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.09.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Blood vessels; Decoding; Deep-shallow hierarchical feature fusion (dsHFF); dual local attention (DLA); Edge detection; Feature extraction; global transformer (GT); Image edge detection; Image Processing, Computer-Assisted; Image segmentation; medical image analysis; Medical imaging; Retinal images; retinal vessel segmentation; Retinal vessels; Retinal Vessels - diagnostic imaging; Semantics; Transformers

Clinically, retinal vessel segmentation is a significant step in the diagnosis of fundus diseases. However, recent methods generally neglect the difference of semantic information between deep and shallow features, which fail to capture the global and local characterizations in fundus images simulta...