A Recursive Attractor Network for Long-Form Sound Source Localization and Identity Tracking with a Variable Number of Sources

Du J, Wu X, Qu T. A Recursive Attractor Network for Long-Form Sound Source Localization and Identity Tracking with a Variable Number of Sources. In: Convention Paper. Audio Engineering Society; 2026.pp. 10271:1-. Available from: https://aes.org/publications/elibrary-page/?id=23183

@inproceedings{Du2026_23183,
  author = {Du, Jiaqi and Wu, Xihong and Qu, Tianshu},
  title = {{A Recursive Attractor Network for Long-Form Sound Source Localization and Identity Tracking with a Variable Number of Sources}},
  pages = {10271:1-},
  year = {2026},
  month = may,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=23183}
}

TY  - CPAPER
TI  - A Recursive Attractor Network for Long-Form Sound Source Localization and Identity Tracking with a Variable Number of Sources
AU  - Du, Jiaqi
AU  - Wu, Xihong
AU  - Qu, Tianshu
SP  - 1
M1  - Paper 10271
PY  - 2026
DA  - 2026/05/28
UR  - https://aes.org/publications/elibrary-page/?id=23183
PB  - Audio Engineering Society
LA  - en
AB  - 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 localization and 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.
ER  -