Estimation of propagation delays using orientation histograms for anechoic blind source separation

Blind source separation (BSS) is one of the most active research areas in neural networks. Previously, we proposed a BSS algorithm which exploited orientation histograms in a joint distribution of observed mixtures. Although the algorithm worked well for simultaneous mixtures of sources, interaural source propagation delays almost always exist in natural environments. In this paper-we extend the algorithm by introducing a bank of band-pass filters and various delays for mixture signals to estimate source propagation delays. It will be shown that the extended algorithm can estimate source delays and separate mixtures satisfactorily. Since our algorithm does not contain learning and does not suffer from arbitrariness in permutation of recovered sources as ICA algorithms do, it could be much simpler and easier to construct a real-time BSS system.