Section: New Results

Detection, Localization and Tracking of Multiple Audio Sources

We addressed the problem of online detection, localization and tracking of multiple moving speakers in reverberant environments [36]. The work has the following contributions. We used the direct-path relative transfer function (DP-RTF), an inter-channel feature that encodes acoustic information robust against reverberation, and we proposed an online algorithm well suited for estimating DP-RTFs associated with moving audio sources. Another crucial ingredient of the proposed method is its ability to properly assign DP-RTFs to audio-source directions. Towards this goal, we adopted a maximum-likelihood formulation and we proposed to use the exponentiated gradient (EG) to efficiently update source-direction estimates starting from their currently available values. The problem of multiple-speaker tracking is computationally intractable because the number of possible associations between observed source directions and physical speakers grows exponentially with time. We adopt a Bayesian framework and we proposed two variational approximations of the posterior filtering distributions associated with multiple speaker tracking, as well as two efficient variational expectation maximization (VEM) solvers [41], [37]. The proposed online localization and tracking methods were thoroughly evaluated using two datasets that contain recordings performed in real environments.

Websites:

https://team.inria.fr/perception/research/audiotrack-vonm/

https://team.inria.fr/perception/research/multi-speaker-tracking/.