计算机工程与应用
計算機工程與應用
계산궤공정여응용
COMPUTER ENGINEERING AND APPLICATIONS
2013年
8期
217-221
,共5页
识别%片上系统(SOC)%现场可编程门阵列(FPGA)%隐马尔可夫模型(HMM)%矢量量化(VQ)
識彆%片上繫統(SOC)%現場可編程門陣列(FPGA)%隱馬爾可伕模型(HMM)%矢量量化(VQ)
식별%편상계통(SOC)%현장가편정문진렬(FPGA)%은마이가부모형(HMM)%시량양화(VQ)
recognition%System on a Chip(SOC)%Field Programmable Gate Array(FPGA)%Hidden Markov Model (HMM)%Vector Quantization(VQ)
随着微电子技术的高速发展,基于片上系统SOC的关键词识别系统的研究已成为当前语音处理领域的研究热点和难点.运用Xilinx公司ViterxII Pro开发板作为硬件平台,结合ISE10.1集成开发环境,完成了语音帧输出、MFCC、VQ和HMM等子模块的设计;提出了一种语音帧压缩模块架构,有效实现了语音帧信息到VQ标号序列的压缩,实现了由语音帧压缩模块和HMM模块构建的FPGA关键词识别系统.仿真实验结果表明,该系统具有较高的识别率和实时性,为关键词识别系统的FPGA硬件电路的实现研究提供了实例.
隨著微電子技術的高速髮展,基于片上繫統SOC的關鍵詞識彆繫統的研究已成為噹前語音處理領域的研究熱點和難點.運用Xilinx公司ViterxII Pro開髮闆作為硬件平檯,結閤ISE10.1集成開髮環境,完成瞭語音幀輸齣、MFCC、VQ和HMM等子模塊的設計;提齣瞭一種語音幀壓縮模塊架構,有效實現瞭語音幀信息到VQ標號序列的壓縮,實現瞭由語音幀壓縮模塊和HMM模塊構建的FPGA關鍵詞識彆繫統.倣真實驗結果錶明,該繫統具有較高的識彆率和實時性,為關鍵詞識彆繫統的FPGA硬件電路的實現研究提供瞭實例.
수착미전자기술적고속발전,기우편상계통SOC적관건사식별계통적연구이성위당전어음처리영역적연구열점화난점.운용Xilinx공사ViterxII Pro개발판작위경건평태,결합ISE10.1집성개발배경,완성료어음정수출、MFCC、VQ화HMM등자모괴적설계;제출료일충어음정압축모괴가구,유효실현료어음정신식도VQ표호서렬적압축,실현료유어음정압축모괴화HMM모괴구건적FPGA관건사식별계통.방진실험결과표명,해계통구유교고적식별솔화실시성,위관건사식별계통적FPGA경건전로적실현연구제공료실례.
With the rapid development of microelectronics, the research of keyword recognition system based on SOC(System on a Chip)has become a hot and difficult topic of speech processing field. This paper completes the module design of speech frame output, Mel Frequency Cepstrum Coefficient(MFCC), Vector Quantization(VQ)and Hidden Markov Model(HMM), using ViterxII Pro board by Xilinx Inc combined with ISE10. 1 integrated development environment. A new structure of speech frame compression module is proposed, which effectively achieves the compression from the speech frame information to VQ label sequence. The Field Programmable Gate Array(FPGA)keyword recognition system is also constructed based on the speech frame compression module and HMM module. The experimental results show that the system has a high recognition rate and good real-time performance. The paper provides an example for studying keyword recognition system based on FPGA.
