Modality Mixer for Multi-modal Action Recognition

TL;DR

The paper introduces M-Mixer, a novel multi-modal action recognition network that effectively leverages complementary modalities and temporal context, outperforming existing methods on multiple benchmark datasets.

Contribution

It proposes the M-Mixer network with the Multi-modal Contextualization Unit (MCU), a new recurrent component that encodes temporal and cross-modal information for improved action recognition.

Findings

Outperforms state-of-the-art on NTU RGB+D 60, NTU RGB+D 120, NW-UCLA datasets.

Demonstrates the effectiveness of the MCU in encoding multi-modal temporal features.

Provides comprehensive ablation studies validating the approach.

Abstract

In multi-modal action recognition, it is important to consider not only the complementary nature of different modalities but also global action content. In this paper, we propose a novel network, named Modality Mixer (M-Mixer) network, to leverage complementary information across modalities and temporal context of an action for multi-modal action recognition. We also introduce a simple yet effective recurrent unit, called Multi-modal Contextualization Unit (MCU), which is a core component of M-Mixer. Our MCU temporally encodes a sequence of one modality (e.g., RGB) with action content features of other modalities (e.g., depth, IR). This process encourages M-Mixer to exploit global action content and also to supplement complementary information of other modalities. As a result, our proposed method outperforms state-of-the-art methods on NTU RGB+D 60, NTU RGB+D 120, and NW-UCLA datasets. Moreover, we demonstrate the effectiveness of M-Mixer by conducting comprehensive ablation studies.