Text-Independent Speaker Iden tification using Soft Channel Selection in a Multi-Microphone Environment

With the aim of improving speaker identification in a multi-microphone environment, we develop a text-independent speaker identification system. It incorporates soft channel selection before the combination of the identification results obtained by multiple microphones. The results demonstrate that the proposed system achieves high classification accuracy, thereby providing a speech interface for a wide range of potential hands-free applications in a ubiquitous environment.