MarginNCE: Robust Sound Localization with a Negative Margin

TL;DR

This paper introduces MarginNCE, a contrastive learning method with a negative margin that improves robustness in sound source localization amidst noisy audio-visual data, outperforming existing methods.

Contribution

The paper proposes a novel contrastive loss modification using a negative margin to handle noisy correspondences in sound localization tasks.

Findings

MarginNCE achieves on-par or better performance than state-of-the-art methods.

Introducing a negative margin consistently improves existing contrastive approaches.

The approach effectively mitigates noise in audio-visual correspondence for localization.

Abstract

The goal of this work is to localize sound sources in visual scenes with a self-supervised approach. Contrastive learning in the context of sound source localization leverages the natural correspondence between audio and visual signals where the audio-visual pairs from the same source are assumed as positive, while randomly selected pairs are negatives. However, this approach brings in noisy correspondences; for example, positive audio and visual pair signals that may be unrelated to each other, or negative pairs that may contain semantically similar samples to the positive one. Our key contribution in this work is to show that using a less strict decision boundary in contrastive learning can alleviate the effect of noisy correspondences in sound source localization. We propose a simple yet effective approach by slightly modifying the contrastive loss with a negative margin. Extensive experimental results show that our approach gives on-par or better performance than the state-of-the-art methods. Furthermore, we demonstrate that the introduction of a negative margin to existing methods results in a consistent improvement in performance.