Energy-efficient MFCC extraction architecture in mixed-signal domain for automatic speech recognition | IEEE Conference Publication | IEEE Xplore