DOA-Aware Audio-Visual Self-Supervised Learning for Sound Event Localization and Detection

TL;DR

This paper introduces a self-supervised pretraining method for sound event localization and detection using audio-visual data, improving performance with less annotated data by jointly learning spatial audio and visual features.

Contribution

It proposes a novel DOA-aware self-supervised learning approach that leverages unannotated audio-visual recordings to enhance sound event localization and detection.

Findings

Reduced SELD error score from 36.4 to 34.9 on DCASE2022 dataset.

Demonstrated effectiveness of self-supervised pretraining with 100 hours of unannotated data.

Improved localization and detection accuracy over traditional supervised methods.

Abstract

This paper describes sound event localization and detection (SELD) for spatial audio recordings captured by firstorder ambisonics (FOA) microphones. In this task, one may train a deep neural network (DNN) using FOA data annotated with the classes and directions of arrival (DOAs) of sound events. However, the performance of this approach is severely bounded by the amount of annotated data. To overcome this limitation, we propose a novel method of pretraining the feature extraction part of the DNN in a self-supervised manner. We use spatial audio-visual recordings abundantly available as virtual reality contents. Assuming that sound objects are concurrently observed by the FOA microphones and the omni-directional camera, we jointly train audio and visual encoders with contrastive learning such that the audio and visual embeddings of the same recording and DOA are made close. A key feature of our method is that the DOA-wise audio embeddings are jointly extracted from the raw audio data, while the DOA-wise visual embeddings are separately extracted from the local visual crops centered on the corresponding DOA. This encourages the latent features of the audio encoder to represent both the classes and DOAs of sound events. The experiment using the DCASE2022 Task 3 dataset of 20 hours shows non-annotated audio-visual recordings of 100 hours reduced the error score of SELD from 36.4 pts to 34.9 pts.