Scalable source localization with multichannel α-stable distributions. Fontaine, M., Vanwynsberghe, C., Liutkus, A., & Badeau, R. In 2017 25th European Signal Processing Conference (EUSIPCO), pages 11-15, Aug, 2017.
doi  abstract   bibtex   
In this paper, we focus on the problem of sound source localization and we propose a technique that exploits the known and arbitrary geometry of the microphone array. While most probabilistic techniques presented in the past rely on Gaussian models, we go further in this direction and detail a method for source localization that is based on the recently proposed α-stable harmonizable processes. They include Cauchy and Gaussian as special cases and their remarkable feature is to allow a simple modeling of impulsive and real world sounds with few parameters. The approach we present builds on the classical convolutive mixing model and has the particularities of requiring going through the data only once, to also work in the underdetermined case of more sources than microphones and to allow massively parallelizable implementations operating in the time-frequency domain. We show that the method yields interesting performance for acoustic imaging in realistic simulations.
@InProceedings{8081159,
  author = {M. Fontaine and C. Vanwynsberghe and A. Liutkus and R. Badeau},
  booktitle = {2017 25th European Signal Processing Conference (EUSIPCO)},
  title = {Scalable source localization with multichannel α-stable distributions},
  year = {2017},
  pages = {11-15},
  abstract = {In this paper, we focus on the problem of sound source localization and we propose a technique that exploits the known and arbitrary geometry of the microphone array. While most probabilistic techniques presented in the past rely on Gaussian models, we go further in this direction and detail a method for source localization that is based on the recently proposed α-stable harmonizable processes. They include Cauchy and Gaussian as special cases and their remarkable feature is to allow a simple modeling of impulsive and real world sounds with few parameters. The approach we present builds on the classical convolutive mixing model and has the particularities of requiring going through the data only once, to also work in the underdetermined case of more sources than microphones and to allow massively parallelizable implementations operating in the time-frequency domain. We show that the method yields interesting performance for acoustic imaging in realistic simulations.},
  keywords = {audio signal processing;blind source separation;convolution;geometry;microphone arrays;statistical distributions;time-frequency analysis;Wiener filters;multichannel α-stable distributions;sound source localization;microphone array;probabilistic techniques;Gaussian models;α-stable harmonizable processes;classical convolutive mixing model;microphones;scalable source localization;acoustic imaging;Direction-of-arrival estimation;Time-frequency analysis;Microphone arrays;Acoustics;Computational modeling;source localization;acoustic modeling;α-stable random variables;spectral measure;sketching},
  doi = {10.23919/EUSIPCO.2017.8081159},
  issn = {2076-1465},
  month = {Aug},
}

Downloads: 0