Dynamic Inter-Class Confusion-Aware Encoder for Audio-Visual Fusion in Human Activity Recognition

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Cong, Kaixuan, Wang, Yifan, Xue, Rongkun, Jiang, Yuyang, Feng, Yiming, Yang, Jing
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918090632069120
author Cong, Kaixuan
Wang, Yifan
Xue, Rongkun
Jiang, Yuyang
Feng, Yiming
Yang, Jing
author_facet Cong, Kaixuan
Wang, Yifan
Xue, Rongkun
Jiang, Yuyang
Feng, Yiming
Yang, Jing
contents Humans do not understand individual events in isolation; rather, they generalize concepts within classes and compare them to others. Existing audio-video pre-training paradigms only focus on the alignment of the overall audio-video modalities, without considering the reinforcement of distinguishing easily confused classes through cognitive induction and contrast during training. This paper proposes the Dynamic Inter-Class Confusion-Aware Encoder (DICCAE), an encoder that aligns audio-video representations at a fine-grained, category-level. DICCAE addresses category confusion by dynamically adjusting the confusion loss based on inter-class confusion degrees, thereby enhancing the model's ability to distinguish between similar activities. To further extend the application of DICCAE, we also introduce a novel training framework that incorporates both audio and video modalities, as well as their fusion. To mitigate the scarcity of audio-video data in the human activity recognition task, we propose a cluster-guided audio-video self-supervised pre-training strategy for DICCAE. DICCAE achieves near state-of-the-art performance on the VGGSound dataset, with a top-1 accuracy of 65.5%. We further evaluate its feature representation quality through extensive ablation studies, validating the necessity of each module.
format Preprint
id arxiv_https___arxiv_org_abs_2507_09323
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dynamic Inter-Class Confusion-Aware Encoder for Audio-Visual Fusion in Human Activity Recognition
Cong, Kaixuan
Wang, Yifan
Xue, Rongkun
Jiang, Yuyang
Feng, Yiming
Yang, Jing
Computer Vision and Pattern Recognition
Humans do not understand individual events in isolation; rather, they generalize concepts within classes and compare them to others. Existing audio-video pre-training paradigms only focus on the alignment of the overall audio-video modalities, without considering the reinforcement of distinguishing easily confused classes through cognitive induction and contrast during training. This paper proposes the Dynamic Inter-Class Confusion-Aware Encoder (DICCAE), an encoder that aligns audio-video representations at a fine-grained, category-level. DICCAE addresses category confusion by dynamically adjusting the confusion loss based on inter-class confusion degrees, thereby enhancing the model's ability to distinguish between similar activities. To further extend the application of DICCAE, we also introduce a novel training framework that incorporates both audio and video modalities, as well as their fusion. To mitigate the scarcity of audio-video data in the human activity recognition task, we propose a cluster-guided audio-video self-supervised pre-training strategy for DICCAE. DICCAE achieves near state-of-the-art performance on the VGGSound dataset, with a top-1 accuracy of 65.5%. We further evaluate its feature representation quality through extensive ablation studies, validating the necessity of each module.
title Dynamic Inter-Class Confusion-Aware Encoder for Audio-Visual Fusion in Human Activity Recognition
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.09323