Date Presented:

28-30 May

摘要:

Sound source localization and identity tracking are fundamental tasks in acoustic scene analysis, enabling machines to determine what, where, and when sound events occur. While deep attractor-based networks have demonstrated improved performance under an unknown number of sources, maintaining continuous source tracking over longform audio remains challenging due to memory limitations and permutation ambiguities across adjacent segments. In this paper, we propose a Recursive Attractor Network (RANet) for long-form sound source localizationand identity tracking with a variable number of sources. RANet explicitly represents attractors as transferable embeddings and recursively propagates them across adjacent audio segments using a LSTM-based model, thereby preserving source identity continuity over time. Experimental results on simulated datasets demonstrate that RANet achieves robust long-form localization and consistent source identity tracking, outperforming baseline approaches. 

访问链接