ANA 发表于 2025-3-25 04:19:23

http://reply.papertrans.cn/88/8741/874021/874021_21.png

glomeruli 发表于 2025-3-25 08:31:51

Takaaki Hori,Atsushi Nakamuraion not only in telecommunication, starting from telephone exchanges through ATM switches and IP routers to optical cross-connect systems and optical packet switches, but also in other areas of knowledge like computation and control. Theory of switching fabrics becomes also a part of applied mathema

食料 发表于 2025-3-25 14:58:53

http://reply.papertrans.cn/88/8741/874021/874021_23.png

不易燃 发表于 2025-3-25 19:35:28

1932-121X he Weighted Finite-State Transducer (WFST) approach. The decoding process for speech recognition is viewed as a search problem whose goal is to find a sequence of words that best matches an input speech signal. Since this process becomes computationally more expensive as the system vocabulary size i

NATTY 发表于 2025-3-25 20:08:42

http://reply.papertrans.cn/88/8741/874021/874021_25.png

Grievance 发表于 2025-3-26 01:09:08

Summary and Perspective,oding with the WFST. Finally, we described two dynamic decoders with on-the-fly composition of multiple WFSTs, which are more tractable for practical speech applications than the basic approach with a fully composed WFST.

亲爱 发表于 2025-3-26 04:54:27

Brief Overview of Speech Recognition,owever, since the technology is wide ranging and progressing day by day, this chapter focuses specifically on a standard approach to large vocabulary continuous speech recognition (LVCSR), which covers continuous-density hidden Markov models, context-dependent phone modeling, .-gram language models,

门窗的侧柱 发表于 2025-3-26 09:10:38

http://reply.papertrans.cn/88/8741/874021/874021_28.png

终端 发表于 2025-3-26 14:01:50

http://reply.papertrans.cn/88/8741/874021/874021_29.png

危机 发表于 2025-3-26 17:53:02

Summary and Perspective,tals of large-vocabulary continuous-speech recognition (LVCSR). Specifically, we focused on the decoding problem in LVCSR and discussed why the WFST-based approach improved search efficiency in the decoding process. We also presented concrete algorithms for constructing a fully composed WFST and dec
页: 1 2 [3] 4
查看完整版本: Titlebook: Speech Recognition Algorithms Using Weighted Finite-State Transducers; Takaaki Hori,Atsushi Nakamura Book 2013 Springer Nature Switzerland