A speech enhancement method based on Kalman filtering

In this paper,We present the enhancement of speech signals polluted by AddingNoise through a filtering method from the dynamic system Kalman Filtering. The AddingNoises to be processed are broadband white noice and color noice. The noise-contaminatedspeech signals are only accessible low signal-noise-ratio signals,without referencing to noisesource,after enhancement,the speech signal's SNR achieves an improvement of approximately 7-10 dB.