Abstract
this paper proposed a new feature enhancement method based on Computational Auditory Scene Analysis (CASA) for dealing with the issue of robust speech recognition. Being different from traditional speech enhancement approaches which focus on clearing the noise from speech mixture as clean as possible, CASA aims to make use of Ideal Binary Mask (IBM) providing sufficient information for human speech recognition. Our experiment on standard Mandarin 863 test corpus plus Noise92 shows that the feature compensation based on CASA can comparatively achieve better improvement for robust speech recognition, particularly bringing about absolute from 20% to 40% improvement at-5dB SNR. Our experiment suggests that CASA method opens a new significant solution to dealing with the issue of robust speech recognition.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
