Denoising Deep Neural Networks Based Voice Activity Detection

TL;DR

This paper introduces a denoising deep neural network (DDNN) for voice activity detection that improves performance over previous deep belief network (DBN) methods by effectively leveraging denoising pre-training and supervised fine-tuning.

Contribution

The paper proposes a novel DDNN-based VAD that uses denoising pre-training to enhance deep layer performance, outperforming existing DBN-based VAD methods.

Findings

DDNN-based VAD outperforms DBN-based VAD.

Deep layers in DDNN show significant performance improvements.

Denoising pre-training enhances deep neural network effectiveness.

Abstract

Recently, the deep-belief-networks (DBN) based voice activity detection (VAD) has been proposed. It is powerful in fusing the advantages of multiple features, and achieves the state-of-the-art performance. However, the deep layers of the DBN-based VAD do not show an apparent superiority to the shallower layers. In this paper, we propose a denoising-deep-neural-network (DDNN) based VAD to address the aforementioned problem. Specifically, we pre-train a deep neural network in a special unsupervised denoising greedy layer-wise mode, and then fine-tune the whole network in a supervised way by the common back-propagation algorithm. In the pre-training phase, we take the noisy speech signals as the visible layer and try to extract a new feature that minimizes the reconstruction cross-entropy loss between the noisy speech signals and its corresponding clean speech signals. Experimental results show that the proposed DDNN-based VAD not only outperforms the DBN-based VAD but also shows an apparent performance improvement of the deep layers over shallower layers.