Vision-language pre-trained models such as CLIP have shown promising cross-scene generalization for visual anomaly detection, but they remain vulnerable to false positives under complex backgrounds and often produce fragmented anomaly regions. In this paper, we study source-supervised zero-shot anomaly detection, where training uses labeled source-domain data and testing is conducted directly on unseen target domains. To address these issues, we propose BSR-CLIP, a background-calibrated and structure-aware vision-language anomaly detection framework that improves the reliability and spatial consistency of anomaly responses. Specifically, we introduce a Structure-Aware Response Correlation Spreading module (RCS), which leverages intermediate-layer self-attention to propagate anomaly responses along structurally related regions, and a Spatial Response Refinement module (SRR), which suppresses noise through intensity-gated adaptive smoothing. Experiments on multiple benchmarks show that BSR-CLIP achieves superior performance under direct source-to-target transfer, including gains of +0.5 I-AP on VisA, +2.5 P-PRO on Kvasir, and +1.0 in both image-level AUROC and AP on Br35H. In addition, BSR-CLIP demonstrates strong robustness in few-shot settings.

Medical image segmentation faces two challenges: First, the contour structures of salient objects and background details vary significantly in medical images of different modalities. Second, the salient and non-salient objects often exhibit misleading co-occurrence phenomena. To overcome these challenges, we propose ASC-Seg. It adopts the Adaptive Structure Encoder (ASE) to capture the global structure of salient objects via parallel Mamba, and uses deformable learning to adaptively locate contour details in complex backgrounds in medical images of different modalities. Plus, our Cross-scale Noise Suppression Decoder (CNSD) uses noise suppression strategy and efficient attention mechanism to suppress non-salient regions and highlight critical salient regions, thereby distinguishing salient and nonsalient objects. Extensive experiments of ASC-Seg on 5 medical image datasets verify its superior performance. Our code is available on https://github.com/SubmissionPaper2025/ASC-Seg

Efficient and effective perception and aggregation mechanisms are crucial in medical image segmentation, especially in scenarios with limited computing resources. However, existing methods are usually accompanied by high computational costs. At the same time, the ubiquitous co-occurrence phenomenon makes it difficult for the model to effectively distinguish target features from interfering background information. To address the above problems, this paper proposes an LSDF-UNet segmentation model. First, a “looking at the big-focusing on the small” dynamic fusion mechanism is adopted to achieve lightweight and efficient feature enhancement under linear complexity. Secondly, a dual-size patch frequency aware module (DPFA) is designed, combined with a frequency-aware block (FAB) and a dual-scale patch partitioning strategy to separate high-frequency details and low-frequency contours in the frequency domain, suppress cooccurrence noise, and significantly improve boundary discrimination. Through extensive experiments on three benchmark medical image datasets, it is demonstrated that our method achieves state-of-the-art performance and effectiveness. Our released code is available at https://github.com/Submission2025/LSDF-UNet.

Medical image segmentation is crucial for clinical decision-making, treatment planning, and disease tracking. Nonetheless, it confronts two significant challenges: the presence of ''soft boundaries'' between the foreground and background exacerbated by poor illumination and low contrast, and the misleading co-occurrence of salient and non-salient objects during the training phase, which complicates the model's accuracy in distinguishing relevant features. To overcome these challenges, we introduce RoDeCon-Net, a novel framework engineered to enhance medical image segmentation. RoDeCon-Net incorporates a Feature Decoupling Unit (FDU) that dynamically separates encoded features into foreground, background, and uncertain regions, using advanced attention mechanisms to refine feature distinction and reduce uncertainty. Additionally, our Contrast-driven Feature Alignment Unit (CFAU) and Cross-layer Feature Cascade Unit (CFCU) synergize to reinforce feature contrasts and promote effective multi-level feature fusion, thus improving the detection of salient objects amidst complex backgrounds and handling various object scales within images. Comprehensive evaluations of RoDeCon-Net on five diverse medical image datasets validate its superior performance and versatility, showcasing its potential to set new benchmarks in medical image segmentation. Our code is available on https://github.com/ILoveACM-MM/RoDeCon-Net.