AI RESEARCH
IsoNet: Spatially-aware audio-visual target speech extraction in complex acoustic environments
arXiv CS.LG
•
ArXi:2605.14736v1 Announce Type: cross Target speech extraction remains difficult for compact devices because monaural neural models lack spatial evidence and classical beamformers lose resolving power when the microaperture is only a few centimetres. We present IsoNet, a user-selectable audio-visual target speech extraction system for a compact 4-microarray. IsoNet combines complex multi-channel STFT features, GCC-PHAT spatial cues, face-conditioned visual embeddings, and auxiliary direction-of-arrival supervision inside a U-Net mask estimation network.