Hearing and Seeing Through CLIP: A Framework for Self-Supervised Sound Source Localization

TL;DR

This paper extends CLIP to sound source localization by mapping audio into tokens compatible with CLIP, enabling self-supervised, zero-shot sound localization with improved accuracy and generalization.

Contribution

It introduces a novel framework that aligns audio embeddings with visual features using CLIP without explicit text input, enhancing sound source localization.

Findings

Outperforms state-of-the-art methods across five tasks

Achieves strong zero-shot generalization

Produces more complete and compact localization results

Abstract

Large-scale vision-language models demonstrate strong multimodal alignment and generalization across diverse tasks. Among them, CLIP stands out as one of the most successful approaches. In this work, we extend the application of CLIP to sound source localization, proposing a self-supervised method operates without explicit text input. We introduce a framework that maps audios into tokens compatible with CLIP's text encoder, producing audio-driven embeddings. These embeddings are used to generate sounding region masks, from which visual features are extracted and aligned with the audio embeddings through a contrastive audio-visual correspondence objective. Our findings show that alignment knowledge of pre-trained multimodal foundation model enables our method to generate more complete and compact localization for sounding objects. We further propose an LLM-guided extension that distills object-aware audio-visual scene understanding into the model during training to enhance alignment. Extensive experiments across five diverse tasks demonstrate that our method, in all variants, outperforms state-of-the-art approaches and achieves strong generalization in zero-shot settings.