MVANet: Multi-Stage Video Attention Network for Sound Event Localization and Detection with Source Distance Estimation

TL;DR

This paper introduces MVANet, a multi-stage video attention network for 3D sound event localization and detection, including source distance estimation, achieving state-of-the-art results in the DCASE 2024 Challenge.

Contribution

The paper presents a novel multi-stage audio-visual network with a new output representation for combined DOA and distance estimation in 3D SELD.

Findings

Outperforms top methods in DCASE 2024 Challenge

Effective use of multi-stage audio features and data augmentation

Accurate source distance estimation in 3D sound localization

Abstract

Sound event localization and detection with source distance estimation (3D SELD) involves not only identifying the sound category and its direction-of-arrival (DOA) but also predicting the source's distance, aiming to provide full information about the sound position. This paper proposes a multi-stage video attention network (MVANet) for audio-visual (AV) 3D SELD. Multi-stage audio features are used to adaptively capture the spatial information of sound sources in videos. We propose a novel output representation that combines the DOA with distance of sound sources by calculating the real Cartesian coordinates to address the newly introduced source distance estimation (SDE) task in the Detection and Classification of Acoustic Scenes and Events (DCASE) 2024 Challenge. We also employ a variety of effective data augmentation and pre-training methods. Experimental results on the STARSS23 dataset have proven the effectiveness of our proposed MVANet. By integrating the aforementioned techniques, our system outperforms the top-ranked method we used in the AV 3D SELD task of the DCASE 2024 Challenge without model ensemble. The code will be made publicly available in the future.