Towards Green ASR: Lossless 4-bit Quantization of a Hybrid TDNN System on the 300-hr Swithboard Corpus.
Junhao XuShoukang HuXunying LiuHelen MengPublished in: INTERSPEECH (2022)
Keyphrases
- spontaneous speech
- uniform quantization
- automatic speech recognition
- adaptive quantization
- image compression
- successive approximation
- speech recognition
- manually annotated
- lossless compression
- lossless image coding
- human machine interaction
- low resolution
- high resolution
- entropy coding
- wavelet transform
- high quality
- image processing