An improved vector quantization method using deep neural network |
| |
| Affiliation: | 1. Department of Electronic Engineering, Shanghai Jiao Tong University, Shanghai 200240, PR China;2. Air Control and Navigation Institution, Air Force Engineering University, Xian 710000, China;1. EHF Key Lab of Science, University of Electronic Science and Technology of China (UESTC), Chengdu 611731, China;2. Institute of Electromagnetics and Acoustics & Department of Electronic Science, Xiamen University, Xiamen 361005, China;3. Shenzhen Research Institute of Xiamen University, Shenzhen 518057, China;4. Department of Electrical and Computer Engineering, Duke University, Durham, NC 27708, USA;1. College of Computer, National University of Defense Technology, 410073 Changsha, Hunan, PR China;2. Institute for Communication Technologies and Embedded Systems, RWTH Aachen University, 52056 Aachen, Germany;3. School of Computer Engineering, Nanyang Technological University, Singapore;1. Aeronautical Key Laboratory for Digital Manufacturing Technologies, AVIC Beijing Aeronautical Manufacturing Technology Research Institute, Baliqiao, Chaoyang District, Beijing 100024, China;2. Key Laboratory of Inertial Science and Technology for National Defence, School of Instrument Science and Opto-Electronics Engineering, Beihang University, 37 Xueyuan Road, Haidian District, Beijing 100191, China;1. Sichuan University, Image Information Institute, College of Electronics and Information Engineering, No. 29 Jiuyanqiao Wangjiang Road, Chengdu 610064, China;2. Northwest University for Nationalities, College of Electrical Engineering, Lanzhou 730030, China |
| |
| Abstract: | To address the challenging problem of vector quantization (VQ) for high dimensional vector using large coding bits, this work proposes a novel deep neural network (DNN) based VQ method. This method uses a k-means based vector quantizer as an encoder and a DNN as a decoder. The decoder is initialized by the decoder network of deep auto-encoder, fed with the codes provided by the k-means based vector quantizer, and trained to minimize the coding error of VQ system. Experiments on speech spectrogram coding demonstrate that, compared with the k-means based method and a recently introduced DNN-based method, the proposed method significantly reduces the coding error. Furthermore, in the experiments of coding multi-frame speech spectrogram, the proposed method achieves about 11% relative gain over the k-means based method in terms of segmental signal to noise ratio (SegSNR). |
| |
| Keywords: | Deep neural network Vector quantization Auto-encoder Binary coding |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
