arXiv 2601.216482026-02-09cs.CVcs.CYcs.HC

CAF-Mamba：基于Mamba的跨模态自适应注意力融合用于多模态抑郁症检测

CAF-Mamba: Mamba-Based Cross-Modal Adaptive Attention Fusion for Multimodal Depression Detection

Neuro-Information Technology Group (NIT) IIKT, Otto von Guericke University Magdeburg Magdeburg, Germany（奥托·冯·格里克大学马格德堡分校神经信息技术小组（NIT）IIKT）

Bowen Zhou, Marc-André Fiedler, Ayoub Al-Hamadi

备注：: The paper contains a total of 5 pages and 3 figures. This paper has been accepted for publication in the proceedings of 2026 IEEE ICASSP Conference

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

CAF-Mamba通过基于Mamba的跨模态自适应注意力融合框架，提升多模态抑郁症检测的性能。

摘要：

抑郁症是一种常见的心理健康障碍，严重损害日常功能和生活质量。尽管最近的深度学习方法在抑郁症检测中显示出潜力，但大多数方法依赖于有限的特征类型，忽视了显式的跨模态交互，并使用简单的拼接或静态加权进行融合。为克服这些限制，我们提出了CAF-Mamba，一种基于Mamba的跨模态自适应注意力融合框架。CAF-Mamba不仅显式和隐式地捕捉跨模态交互，还通过模态级注意力机制动态调整模态贡献，从而实现更有效的多模态融合。在两个真实世界基准数据集LMVD和D-Vlog上的实验表明，CAF-Mamba在一致性上优于现有方法，并实现了最先进的性能。我们的代码可在https://github.com/zbw-zhou/CAF-Mamba上获得。

英文摘要：

Depression is a prevalent mental health disorder that severely impairs daily functioning and quality of life. While recent deep learning approaches for depression detection have shown promise, most rely on limited feature types, overlook explicit cross-modal interactions, and employ simple concatenation or static weighting for fusion. To overcome these limitations, we propose CAF-Mamba, a novel Mamba-based cross-modal adaptive attention fusion framework. CAF-Mamba not only captures cross-modal interactions explicitly and implicitly, but also dynamically adjusts modality contributions through a modality-wise attention mechanism, enabling more effective multimodal fusion. Experiments on two in-the-wild benchmark datasets, LMVD and D-Vlog, demonstrate that CAF-Mamba consistently outperforms existing methods and achieves state-of-the-art performance. Our code is available at https://github.com/zbw-zhou/CAF-Mamba.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。