Cascade architecture for noise-robust keyword spotting

Abstract

A method includes receiving, at a first processor of a user device, streaming multi-channel audio captured by an array of microphones, each channel including respective audio features. For each channel, the method also includes processing, by the first processor, using a first stage hotword detector, the respective audio features to determine whether a hotword is detected. When the first stage hotword detector detects the hotword, the method also includes the first processor providing chomped raw audio data to a second processor that processes, using a first noise cleaning algorithm, the chomped raw audio data to generate a clean monophonic audio chomp. The method also includes processing, by the second processor using a second stage hotword detector, the clean monophonic audio chomp to detect the hotword.