Type-2 fuzzy neural network for voice activity detection
Gin-Der Wu, Po-Jen Wu · 2016
Voice activity detection (VAD) is an important classification problem in signal processing. It is easily affected by the noisy environment. To solve this problem, this paper proposes a type-2 fuzzy neural network (T2FNN) for VAD. In handing problems with uncertainties such as noisy data, type-2 fuzzy-systems generally outperform their type-1 counterparts. Hence, type-2 fuzzy-sets are adopted to model the noisy data. By structure learning and parameter learning, T2FNN has high discriminability. Compared with other existing fuzzy neural networks, the novelty of the T2FNN is its consideration of both uncertainty and discriminability. The effectiveness of the T2FNN is demonstrated by speech endpoint detection. Experimental result indicates that T2FNN performs better than the other type-1 fuzzy neural networks.