Loading…
This event has ended. Visit the official site or create your own event on Sched.
Welcome to ICSV26!
Back To Schedule
Wednesday, July 10 • 16:50 - 17:10
A NOISE SUPPRESSION METHOD BASED ON JOINT OBSERVATION OF BONE- AND AIR-CONDUCTED SPEECH SIGNALS BY USE OF AN EXTENSION TYPE UKF

Log in to save this to your schedule, view media, leave feedback and see who's attending!

Feedback form is now closed.
When applying speech recognition to actual circumstances such as inspection and maintenance operations in industrial factories to recording and reporting routines at construction sites, etc. where hand-writing is difficult, some countermeasure methods for surrounding noise are indispensable. In this study, a signal processing method to remove the noise for actual speech signals is proposed by jointly using the measured data of bone- and air-conducted speech signals. The bone- and air-conducted speech signal is generally expressed by a nonlinear model. The extended Kalman filter is very famous as a noise suppression method for nonlinear system model. However, this filter is needs a linearized approximation for the nonlinear model. By using the sample point called Sigma point, unscented Kalman filter (UKF) can be applied to the nonlinear system model without linearized approximation. New type of noise suppression method is proposed by extending the UKF. Although UKF assumes Gaussian noise, the new type extended UKF considers non - Gaussian noise. The effectiveness of the proposed method is confirmed by applying it to bone- and air- conducted speech measured in real environments under the existence of surrounding noise.

Moderators
AR

Annie Ross

Professor, Polytechnique Montreal

Authors

Wednesday July 10, 2019 16:50 - 17:10 EDT
Westmount 3
  T10 Sig. Proc. & nonlin. mthds., RS04 Signal process in ac & vibr