错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Sound Source Localization

  • Jisheng Bai,
  • Dongzhe Zhang,
  • Qingli Yan,
  • Qian Zhu,
  • Jianfeng Chen

摘要

Human perception of sound source location plays a vital role in understanding acoustic environments, enabling spatial awareness, auditory attention, and interaction with surrounding events. Sound source localization, which seeks to computationally replicate this spatial hearing capability, constitutes a critical component of audio analysis systems. This chapter is structured into three interconnected sections: direction of arrival estimation, which focuses on determining the angular position of sound sources; sound source localization, which extends to identify the physical coordinates of sound sources particularly in wireless acoustic sensor networks; and sound event localization and detection, which combines the tasks of identifying sound types while simultaneously localizing their positions. While traditionally dominated by array-signal-processing techniques, this field has recently seen significant advancements through deep learning approaches that have substantially improved performance. Each section of this chapter introduces state-of-the-art deep learning-based sound source localization methods, covering the content such as model architecture, evaluation metrics, and datasets. As a fundamental element of sound perception, this chapter highlights the importance of sound source localization, which serves as the foundation for numerous real-life applications.