An information fusion framework with multi-channel feature concatenation and multi-perspective system combination for the deep-learning-based robust recognition of microphone array speech.
Computer Speech & Language(2017)
摘要
•The early fusion by using multiple beamformings and feature concatenation.•The late fusion of subnets from multiple perspectives.•A simplified and effective MVDR beamforming approach.•Building the best one-pass single DNN system among all submissions to CHiME-3.
更多查看译文
关键词
CHiME challenge,Deep learning,Information fusion,Microphone array,Robust speech recognition
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络