MM-Gesture: Towards Precise Micro-Gesture Recognition through Multimodal Fusion

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Gu, Jihao, Wang, Fei, Li, Kun, Wei, Yanyan, Wu, Zhiliang, Guo, Dan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912520648785920
author Gu, Jihao
Wang, Fei
Li, Kun
Wei, Yanyan
Wu, Zhiliang
Guo, Dan
author_facet Gu, Jihao
Wang, Fei
Li, Kun
Wei, Yanyan
Wu, Zhiliang
Guo, Dan
contents In this paper, we present MM-Gesture, the solution developed by our team HFUT-VUT, which ranked 1st in the micro-gesture classification track of the 3rd MiGA Challenge at IJCAI 2025, achieving superior performance compared to previous state-of-the-art methods. MM-Gesture is a multimodal fusion framework designed specifically for recognizing subtle and short-duration micro-gestures (MGs), integrating complementary cues from joint, limb, RGB video, Taylor-series video, optical-flow video, and depth video modalities. Utilizing PoseConv3D and Video Swin Transformer architectures with a novel modality-weighted ensemble strategy, our method further enhances RGB modality performance through transfer learning pre-trained on the larger MA-52 dataset. Extensive experiments on the iMiGUE benchmark, including ablation studies across different modalities, validate the effectiveness of our proposed approach, achieving a top-1 accuracy of 73.213%. Code is available at: https://github.com/momiji-bit/MM-Gesture.
format Preprint
id arxiv_https___arxiv_org_abs_2507_08344
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MM-Gesture: Towards Precise Micro-Gesture Recognition through Multimodal Fusion
Gu, Jihao
Wang, Fei
Li, Kun
Wei, Yanyan
Wu, Zhiliang
Guo, Dan
Computer Vision and Pattern Recognition
In this paper, we present MM-Gesture, the solution developed by our team HFUT-VUT, which ranked 1st in the micro-gesture classification track of the 3rd MiGA Challenge at IJCAI 2025, achieving superior performance compared to previous state-of-the-art methods. MM-Gesture is a multimodal fusion framework designed specifically for recognizing subtle and short-duration micro-gestures (MGs), integrating complementary cues from joint, limb, RGB video, Taylor-series video, optical-flow video, and depth video modalities. Utilizing PoseConv3D and Video Swin Transformer architectures with a novel modality-weighted ensemble strategy, our method further enhances RGB modality performance through transfer learning pre-trained on the larger MA-52 dataset. Extensive experiments on the iMiGUE benchmark, including ablation studies across different modalities, validate the effectiveness of our proposed approach, achieving a top-1 accuracy of 73.213%. Code is available at: https://github.com/momiji-bit/MM-Gesture.
title MM-Gesture: Towards Precise Micro-Gesture Recognition through Multimodal Fusion
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.08344