Aug 2026
WFDViM: wavelet frequency-decoupled vision Mamba for robust weak-boundary medical image segmentation
WFDViM is proposed, a wavelet frequency-decoupled Vision Mamba framework for weak-boundary medical image segmentation that achieves the lowest mean HD95 and ASSD across all four datasets while requiring 13.06 M parameters and 3.05 GFLOPs.
Chuan-Zhong Huang, Xin-Ao Wu, X. Tan et al.
· Multimedia Systems · 0 citations