Convolutional networks for speech detection

In this paper, we introduce a new framework for speech detection using convolutional networks. We propose a network architecture that can incorporate long and short-term temporal and spectral cor- relations of speech in the detection process. The proposed design is able to address many shortcomings of existing speech detectors in a unified new framework: First, itimproves the robustness of the system to environmental variability while still being fast to evalu- ate. Second, it allows for a framework that is extendable to work under different time-scales for different applications. Finally, it is discriminative and produces reliable estimates of the probability of presence of speech in each frame for a wide variety of noise con- ditions. We propose that the inputs to the system be features that are measures of the true signal-to-noise ratio of a set of frequency bands of the signal. These can be easily and automatically gener- ated by tracking thenoise spectrum online. We present preliminary results on the AURORA database to demonstrate the effectiveness of the detector over conventional Gaussian detectors.