Unsupervised Speech/Non-Speech Detection for Automatic Speech Recognition in Meeting Rooms

Abstract

Abstract. The goal of this work is to provide robust and accurate speech detection for automatic speech recognition (ASR) in meeting room settings. The solution is based on computing long-term modulation spectrum, and examining specific frequency range for dominant speech components to classify speech and non-speech signals for a given audio signal. Manually segmented speech segments, short-term energy, short-term energy and zero-crossing based segmentation techniques, and a recently proposed Multi Layer Perceptron (MLP) classifier system are tested for comparison purposes. Speech recognition evaluations of the segmentation methods are performed on a standard database and tested in conditions where the signal-to-noise ratio (SNR) varies considerably, as in the cases of close-talking headset, lapel, distant microphone array output, and distant microphone. The results reveal that the proposed method is more reliable and less sensitive to mode of signal acquisition and unforeseen conditions.

Investigating automatic measurements of prosodic accommodation and its dynamics in social interaction

“…Speech/silent intervals are detected using a method based on long-term modulation spectrum energy features (Maganti et al, 2007). Detection of syllable nuclei is performed using the method introduced in De Jong and Wempe (2009), which is based on intensity peak detection of voiced segments of speech.…”

Microphone array based classification for security monitoring in unstructured environments

“…The first level is devoted to distinguishing speech from non-speech sound. This task, known under the name of Automatic Speech Detection [10,11,12], has been extensively studied in the literature, since it is basilar to any system requiring speech enhancement, speech recognition and (as in our case) speech classification.…”

“…For example, there exists a vast literature regarding speech discrimination [10,11,12,13], vehicle recognition [14,15,16] and weapon classification [17,18,7]. In addition, due to the maturity of the field there exist several commercial and open-source products that perform these tasks, such as the Halo system 1 and the Sphinx toolkit 2 .…”

Integration of evolutionary computation algorithms and new AUTO-TLBO technique in the speaker clustering stage for speaker diarization of broadcast news

“…The use of the within-class covariance (WCC) matrix to normalize data variances has become widely dispread in the speaker recognition field [41,43]. The need to be normalized for I-vectors which differ from one application to another is due to its representation of a wide range of the speech variability.…”