A low-power, fixed-point, front-end feature extraction for a distributed speech recognition system.
Brian DelaneyNikil JayantMat HansTajana SimunicAndrea AcquavivaPublished in: ICASSP (2002)
Keyphrases
- fixed point
- low power
- feature extraction
- power consumption
- low cost
- high speed
- single chip
- sufficient conditions
- vlsi architecture
- preprocessing
- dynamical systems
- belief propagation
- constraint databases
- logic circuits
- floating point
- digital signal processing
- vlsi circuits
- wavelet transform
- low power consumption
- feature vectors
- power reduction
- message passing
- similarity measure
- database systems
- real time