Adaptive detection of noisy speech using third-order statistics
(EN)
Carayannis, G
(EN)
Rangoussi, M
(EN)
Detection of speech in noisy recordings is challenging, especially when the noise does not follow the usual whiteness, stationarity and high signal-to-noise ratio assumptions. A robust speech detector can affect significantly the performance of several speech-processing tasks, such as endpoint detection, segmentation and finally recognition, if we deal with real life data as opposed to laboratory or controlled environment recordings. The detector proposed in this paper is based on a Gaussianity test that employs third-order cumulants of the data to decide on the binary hypotheses of noise only versus speech plus noise. Speech intervals are detected by exploiting the third-order information present in the speech signal. The detector can tolerate a large family of additive noises thanks to its third-order statistics basis. The sample-adaptive and decision feedback variations proposed here provide the detector with tracking ability with respect to both the time variations of speech and the possible non-stationarity of noise. Experiments carried out using real data recorded in a moving car interior show satisfactory performance of the proposed algorithms down to - 6 dB signal-to-noise ratio.
(EN)
