The proposed UltraLight Vision Mamba UNet (UltraLight VM-UNet) is a highly efficient model for skin lesion segmentation, achieving excellent performance with only 0.049M parameters and 0.060 GFLOPs, which is significantly lower than existing lightweight Vision Mamba models.
The proposed AC-MambaSeg model combines the strengths of CNN and Vision Mamba to effectively capture both local and global features for accurate skin lesion segmentation, further enhanced by advanced components like CBAM and Selective Kernels.
The proposed AD-Net utilizes a dilated convolutional residual network with an attention-based spatial feature enhancement block and a guided decoder strategy to achieve robust and efficient skin lesion segmentation.
SkinMamba, a hybrid architecture combining the strengths of Mamba and CNN, effectively addresses challenges in skin lesion segmentation such as varying lesion sizes and unclear boundaries through cross-scale global modeling and frequency-guided boundary detection.