机译:具有多通道特征串联和多视角系统组合的信息融合框架,用于基于深度学习的麦克风阵列语音鲁棒识别
University of Science and Technology of China, Hefei, Anhui, China;
University of Science and Technology of China, Hefei, Anhui, China;
University of Science and Technology of China, Hefei, Anhui, China;
University of Science and Technology of China, Hefei, Anhui, China;
University of Science and Technology of China, Hefei, Anhui, China;
Georgia Institute of Technology, Atlanta, GA, United States;
CHiME challenge; Deep learning; Information fusion; Microphone array; Robust speech recognition;
机译:通过将多个麦克风阵列处理与位置相关的CMN相结合,实现鲁棒的远程语音识别
机译:通过将多个麦克风阵列处理与位置相关的CMN相结合,实现鲁棒的远程语音识别
机译:麦克风阵列鲁棒语音识别的空间判别功能
机译:分布式麦克风阵列中的多通道功能增强,可在智能房间中实现可靠的远距离语音识别
机译:基于麦克风阵列,视听和帧选择的强大语音处理功能,可实现车载语音识别和内置说话人识别。
机译:在带通滤波调制多流功能根据框架鲁棒语音识别
机译:增强Gmm扬声器识别系统的稳健性,用于低复杂度麦克风阵列的嘈杂和混响语音