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

TL;DR

CAF-Mamba introduces a dynamic, attention-based fusion framework for multimodal depression detection, explicitly modeling cross-modal interactions and adaptively weighting modalities, leading to superior performance on benchmark datasets.

Contribution

It presents a novel Mamba-based adaptive attention fusion method that explicitly captures cross-modal interactions and dynamically adjusts modality importance for depression detection.

Findings

Outperforms existing methods on LMVD and D-Vlog datasets

Achieves state-of-the-art performance in multimodal depression detection

Demonstrates effective cross-modal interaction modeling

Abstract

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.