The paper discusses the study of speaker localization for binaural microphone arrays, which is critical for applications like speech communication, video conferencing, and robot audition. It presents alternatives to current processing stages, inspired by human hearing. This includes the incorporation of an auditory filter bank instead of the short-time Fourier transform, and a new direction of arrival search based on transformed head related transfer function. The proposed methods were validated through simulation and experimental studies, showing favorable comparison with existing methods.
Publication date: 3 Nov 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2310.20238
