Exploring Audio-Visual Information Fusion for Sound Event Localization and Detection In Low-Resource Realistic Scenarios

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Jiang, Ya, Wang, Qing, Du, Jun, Hu, Maocheng, Hu, Pengfei, Liu, Zeyan, Cheng, Shi, Nian, Zhaoxu, Dong, Yuxuan, Cai, Mingqi, Fang, Xin, Lee, Chin-Hui
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909228946423808
author Jiang, Ya
Wang, Qing
Du, Jun
Hu, Maocheng
Hu, Pengfei
Liu, Zeyan
Cheng, Shi
Nian, Zhaoxu
Dong, Yuxuan
Cai, Mingqi
Fang, Xin
Lee, Chin-Hui
author_facet Jiang, Ya
Wang, Qing
Du, Jun
Hu, Maocheng
Hu, Pengfei
Liu, Zeyan
Cheng, Shi
Nian, Zhaoxu
Dong, Yuxuan
Cai, Mingqi
Fang, Xin
Lee, Chin-Hui
contents This study presents an audio-visual information fusion approach to sound event localization and detection (SELD) in low-resource scenarios. We aim at utilizing audio and video modality information through cross-modal learning and multi-modal fusion. First, we propose a cross-modal teacher-student learning (TSL) framework to transfer information from an audio-only teacher model, trained on a rich collection of audio data with multiple data augmentation techniques, to an audio-visual student model trained with only a limited set of multi-modal data. Next, we propose a two-stage audio-visual fusion strategy, consisting of an early feature fusion and a late video-guided decision fusion to exploit synergies between audio and video modalities. Finally, we introduce an innovative video pixel swapping (VPS) technique to extend an audio channel swapping (ACS) method to an audio-visual joint augmentation. Evaluation results on the Detection and Classification of Acoustic Scenes and Events (DCASE) 2023 Challenge data set demonstrate significant improvements in SELD performances. Furthermore, our submission to the SELD task of the DCASE 2023 Challenge ranks first place by effectively integrating the proposed techniques into a model ensemble.
format Preprint
id arxiv_https___arxiv_org_abs_2406_15160
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Exploring Audio-Visual Information Fusion for Sound Event Localization and Detection In Low-Resource Realistic Scenarios
Jiang, Ya
Wang, Qing
Du, Jun
Hu, Maocheng
Hu, Pengfei
Liu, Zeyan
Cheng, Shi
Nian, Zhaoxu
Dong, Yuxuan
Cai, Mingqi
Fang, Xin
Lee, Chin-Hui
Audio and Speech Processing
Signal Processing
This study presents an audio-visual information fusion approach to sound event localization and detection (SELD) in low-resource scenarios. We aim at utilizing audio and video modality information through cross-modal learning and multi-modal fusion. First, we propose a cross-modal teacher-student learning (TSL) framework to transfer information from an audio-only teacher model, trained on a rich collection of audio data with multiple data augmentation techniques, to an audio-visual student model trained with only a limited set of multi-modal data. Next, we propose a two-stage audio-visual fusion strategy, consisting of an early feature fusion and a late video-guided decision fusion to exploit synergies between audio and video modalities. Finally, we introduce an innovative video pixel swapping (VPS) technique to extend an audio channel swapping (ACS) method to an audio-visual joint augmentation. Evaluation results on the Detection and Classification of Acoustic Scenes and Events (DCASE) 2023 Challenge data set demonstrate significant improvements in SELD performances. Furthermore, our submission to the SELD task of the DCASE 2023 Challenge ranks first place by effectively integrating the proposed techniques into a model ensemble.
title Exploring Audio-Visual Information Fusion for Sound Event Localization and Detection In Low-Resource Realistic Scenarios
topic Audio and Speech Processing
Signal Processing
url https://arxiv.org/abs/2406.15160