FCN Approach for Dynamically Locating Multiple Speakers
This work addresses the challenge of accurately and simultaneously localizing and tracking multiple speakers in dynamic scenarios, which is incremental as it builds on existing deep learning methods for speaker localization.
The paper tackled the problem of online multi-speaker localization by using a fully convolutional network to estimate direction of arrival for each time-frequency bin, achieving superior performance over classic and recent deep-learning-based algorithms in simulated and real-life recordings.
In this paper, we present a deep neural network-based online multi-speaker localisation algorithm. Following the W-disjoint orthogonality principle in the spectral domain, each time-frequency (TF) bin is dominated by a single speaker, and hence by a single direction of arrival (DOA). A fully convolutional network is trained with instantaneous spatial features to estimate the DOA for each TF bin. The high resolution classification enables the network to accurately and simultaneously localize and track multiple speakers, both static and dynamic. Elaborated experimental study using both simulated and real-life recordings in static and dynamic scenarios, confirms that the proposed algorithm outperforms both classic and recent deep-learning-based algorithms.