Description of algorithms for Ben-Gurion University Submission to the\n LOCATA challenge

摘要

This paper summarizes the methods used to localize the sources recorded for\nthe LOCalization And TrAcking (LOCATA) challenge. The tasks of stationary\nsources and arrays were considered, i.e., tasks 1 and 2 of the challenge, which\nwere recorded with the Nao robot array, and the Eigenmike array. For both\narrays, direction of arrival (DOA) estimation has been performed with\nmeasurements in the short time Fourier transform domain, and with direct-path\ndominance (DPD) based tests, which aim to identify time-frequency (TF) bins\ndominated by the direct sound. For the recordings with Nao, a DPD test which is\napplied directly to the microphone signals was used. For the Eigenmike\nrecordings, a DPD based test designed for plane-wave density measurements in\nthe spherical harmonics domain was used. After acquiring DOA estimates with TF\nbins that passed the DPD tests, a stage of k-means clustering is performed, to\nassign a final DOA estimate for each speaker.\n