Abstract: Conversational emotion recognition (CER) is an important research topic in human-computer interactions. Although recent advancements in transformer-based cross-modal fusion methods have ...
@inproceedings{Tang2024DRMF, title={DRMF: Degradation-Robust Multi-Modal Image Fusion via Composable Diffusion Prior}, author={Tang, Linfeng and Deng, Yuxin and Yi, Xunpeng and Yan, Qinglong and Yuan, ...
Abstract: A critical challenge for multi-modal Object Re-Identification (ReID) is the effective aggregation of complementary information to mitigate illumination issues. State-of-the-art methods ...
Towards open-ended Video Anomaly Detection (VAD), existing methods often exhibit biased detection when faced with challenging or unseen events and lack interpretability. To address these drawbacks, we ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results