U-Net Based Direct-Path Dominance Test for Robust Direction-of-Arrival Estimation

Abstract

It has been noted that the identification of the time-frequency bins dominated by the contribution from the direct propagation of the target speaker can significantly improve the robustness of the direction-of-arrival estimation. However, the correct extraction of the direct-path sound is challenging especially in adverse environments. In this paper, a U-net based direct-path dominance test method is proposed. Exploiting the efficient segmentation capability of the U-net architecture, the direct-path information can be effectively retrieved from a dedicated multi-task neural network. Moreover, the training and inference of the neural network only need the input of a single microphone, circumventing the problem of array-structure dependence faced by common endto-end deep learning based methods. Simulations demonstrate that significantly higher estimation accuracy can be achieved in high reverberant and low signal-to-noise ratio environments.

Neural-Network-Based Direction-of-Arrival Estimation for Reverberant Speech - The Importance of Energetic, Temporal, and Spatial Information

“…Although this approach has been previously applied to beamforming and source separation [14], it has not been extensively studied for DOA estimation. The study in [21] used the logarithmic magnitude spectrogram of a single-channel noisy signal as the input of a NN that classifies direct-path TF bins, and then estimated the DOA from these bins using the MUSIC algorithm [5] or the steered response power -phase transform (SRP-PHAT) algorithm [22]. While this approach displayed good performance, its disadvantages are the typically large number of the features defined by the short-time Fourier transform (STFT) size, and the long time and large amount of data required for training.…”