Recently, Mamba-based methods have gained popularity in medical image segmentation due to their ability to model long-range dependencies with linear computational complexity. However, current segmentation methods often face challenges such as low contrast, blurred boundaries, and unclear backgrounds in medical images. Considering that perceived objects and features exhibit greater discriminative power in the frequency domain, we propose a novel mamba-based spatial and frequency domain feature fusion network, SFMa-Unet, to address these challenges. Specifically, we designed the Spatial-Frequency Interaction (SFI) module, which leverages the powerful modeling capabilities of Mamba to fuse spatial and frequency domain features, enhancing feature representation. Additionally, we developed a Mamba-based multi-scale feature channel fusion (MFCF) bridge to capture local and global dependencies across different feature scales, further improving the model’s representational capacity. We conduct comprehensive experiments on the ISIC17 and ISIC18 public datasets. Experimental results demonstrate the effectiveness and robustness of SFMa-Unet. Codes are available at https://github.com/RainCh-zyq/SFma-Unet.

The improvement of transferability of adversarial examples is the key property in practical black-box scenarios. Recent research has identified that transferable adversarial examples for video models can be effectively crafted with image models. However, existing studies primarily target single-layer features, overlooking the influence of diverse feature layers. Moreover, they neglect transitions between video frames and fail to fully capture temporal context. In this paper, we introduce an efficient and stable cross-modal attack method termed Feature and Temporal Disruption Attack (FTDA). Our approach caters to both feature space diversity and temporal cues by introducing two innovative modules, i.e., Depth-Aware Feature Fusion Attack (DF2A) and Clip-Based Temporal Fusion Attack (CTFA). Extensive experiments demonstrate that our approach achieves SOTA. Our code is available at https://github.com/xiaopengge2000/FTDA.