Search papers, labs, and topics across Lattice.
The paper introduces CMSA-Net, a novel framework for video polyp segmentation (VPS) designed to address challenges arising from polyp similarity to surrounding tissue and variations in polyp position/scale across frames. CMSA-Net employs a Causal Multi-scale Aggregation (CMA) module to aggregate semantic information from historical frames using causal attention and a Dynamic Multi-source Reference (DMR) strategy to adaptively select informative reference frames. Experiments on the SUN-SEG dataset demonstrate state-of-the-art performance, balancing accuracy and real-time applicability.
Achieve state-of-the-art video polyp segmentation by adaptively selecting informative reference frames and aggregating multi-scale historical features with causal attention.
Video polyp segmentation (VPS) is an important task in computer-aided colonoscopy, as it helps doctors accurately locate and track polyps during examinations. However, VPS remains challenging because polyps often look similar to surrounding mucosa, leading to weak semantic discrimination. In addition, large changes in polyp position and scale across video frames make stable and accurate segmentation difficult. To address these challenges, we propose a robust VPS framework named CMSA-Net. The proposed network introduces a Causal Multi-scale Aggregation (CMA) module to effectively gather semantic information from multiple historical frames at different scales. By using causal attention, CMA ensures that temporal feature propagation follows strict time order, which helps reduce noise and improve feature reliability. Furthermore, we design a Dynamic Multi-source Reference (DMR) strategy that adaptively selects informative and reliable reference frames based on semantic separability and prediction confidence. This strategy provides strong multi-frame guidance while keeping the model efficient for real-time inference. Extensive experiments on the SUN-SEG dataset demonstrate that CMSA-Net achieves state-of-the-art performance, offering a favorable balance between segmentation accuracy and real-time clinical applicability.
