專題進度報告 資工四 B93902054 洪志豪 資工四 B93902076 林宜鴻.

Slides:



Advertisements
Similar presentations
Introduction to C Programming
Advertisements

數位語音處理概論 HW#2-1 HMM Training and Testing
專題研究 WEEK 4 - LIVE DEMO Prof. Lin-Shan Lee TA. Hsiang-Hung Lu,Cheng-Kuan Wei.
Loading Excel Double click the Excel icon on the desktop (if you have this) OR Click on Start All Programs Microsoft Office Microsoft Office Excel 2003.
INSTRUCTOR:Dr.Veton Kepuska STUDENT:Dileep Narayan.Koneru YES/NO RECOGNITION SYSTEM.
The Little man computer
English Shellcode J. Mason, S. Small, F. Monrose, G. MacManus CCS ’09 Presented by: Eugenie Lee EE515/IS523: Security101: Think Like an Adversary.
Chapter 3 Assembly Language: Part 1. Machine language program (in hex notation) from Chapter 2.
XML CS 105. What is XML? XML stands for Extensible Markup Language. XML is a markup language like HTML. XML was designed to describe data. You must define.
語音辨認概論 A Tutorial Example of Using HTK 96/10/18 老師 : 廖元甫 演講者 : 蔡明峰.
Number systems Converting numbers between binary, octal, decimal, hexadecimal (the easy way)
Chapter 2 Software Tools and Assembly Language Syntax.
Introduction to Automatic Speech Recognition
12/13/2007Chia-Ho Ling1 SRILM Language Model Student: Chia-Ho Ling Instructor: Dr. Veton Z. K ë puska.
1M4 speech recognition University of Sheffield M4 speech recognition Martin Karafiát*, Steve Renals, Vincent Wan.
A453 Exemplar Password Program using VBA
August 1, 2000Frame Relay Service Level Definitions 1 Status Current Draft (draft-ietf-frnetmib-frmrelay- service-02.txt) –Addressed ALL issues from last.
Temple University Goals : 1.Down sample 20 khz TIDigits data to 16 khz. 2. Use Down sample data run regression test and Compare results posted in Sphinx-4.
Arnel Fajardo, student (“Hak Seng”)
Presentation by Daniel Whiteley AME department
Exploring an Open Source Automation Framework Implementation.
DSP homework 1 HMM Training and Testing
1M4 speech recognition University of Sheffield M4 speech recognition Vincent Wan, Martin Karafiát.
Results of Tagalog vowel Speech recognition using Continuous HMM Arnel C. Fajardo Ph. D student (Under the supervision of Professor Yoon-Joong Kim)
Modeling Speech using POMDPs In this work we apply a new model, POMPD, in place of the traditional HMM to acoustically model the speech signal. We use.
Aim: How do we multiply or divide complex numbers? Do Now: 1. Multiply: 2. Multiply: 3. Multiply: 6 + 7x + 2x i HW: p.216 # 26,30,32,36,38,40,50,52.
Understanding Allocations Brian Chizever Cognos Corporation.
11/25/2015Slide 1 Scripts are short programs that repeat sequences of SPSS commands. SPSS includes a computer language called Sax Basic for the creation.
Bit-DSP-MicrocontrollerTMS320F2812 Texas Instruments Incorporated European Customer Training Center University of Applied Sciences Zwickau (FH)
專題研究 (4) HDecode_live Prof. Lin-Shan Lee, TA. Yun-Chiao Li 1.
The HTK Book (for HTK Version 3.2.1) Young et al., 2002.
Performance Comparison of Speaker and Emotion Recognition
I’m Thinking of a Number
Math Problem -- ages Making your own variable equation to solve a math riddle.
Section 2.7 Combining Functions Objectives: To add, subtract, multiply and divide functions. Composition of functions.
Fundamental of Databases
Olivier Siohan David Rybach
EET 2259 Unit 13 Strings and File I/O
The Little man computer
QNAP TAS series NAS - Android Dual System
Development Environment
Getting started with the Arxterra software and 3DoT Firmware
Prof. Lin-shan Lee TA. Roy Lu
Repetition Structures Chapter 9
Homework Reading Labs PAL, pp
Combining Like Terms 8th Pre-Algebra.
Assembly Language Programming Part 3
專題研究 week3 Language Model and Decoding
Prof. Lin-shan Lee TA. Lang-Chi Yu
Digital Speech Processing
قانون المنافسة ومنع الاحتكار
CSC 3210 Computer Organization and Programming
التدريب الرياضى إعداد الدكتور طارق صلاح.
C Graphing Functions.
Homework Reading Machine Projects Labs PAL, pp
Electronics II Physics 3620 / 6620
Accessing “Old Patient Chart” from HED
Prof. Lin-shan Lee TA. Po-chun, Hsu
Presentation by Daniel Whiteley AME department
2007 SPEECH PROJECT PRESENTATION
Time Series Prediction with Mixture of Experts
Cheng-Kuan Wei1 , Cheng-Tao Chung1 , Hung-Yi Lee2 and Lin-Shan Lee2
Chapter 4: Repetition Structures: Looping
EET 2259 Unit 13 Strings and File I/O
Visual Recognition of American Sign Language Using Hidden Markov Models 문현구 문현구.
1.3.7 High- and low-level languages and their translators
Prof. Lin-shan Lee TA. Roy Lu
Number systems Converting numbers between binary, octal, decimal, hexadecimal (the easy way)
Potions Project By Francis Mayen.
Presentation transcript:

專題進度報告 資工四 B93902054 洪志豪 資工四 B93902076 林宜鴻

Flow chart

Language Model 1.Normalize(UTF-8) -> normalize_utf8 2.From UTF-8 to Big5 -> iconv 3.parse according to dictory -> textSegment 4.Build language model -> ngram-count

Acoustic Model - HERest 1.meaning:Adjust HMM model 2.command: HERest –C herest.cfg –S train.scp –L /share/data/TrainingLabel –H hmmset.mmf rcd.lst 3.Run more can get higher accuracy

Combine Acoustic and Language Model(1) pre-requirement 1.change the bigram.lm from Big5 to ascii 2.command: /share/program/big52ascii/chenc bigram_Big5.lm > bigram_ascii.lm 3.replace <s> with !=ENTER </s>with !=EXIT A useful command to replace: ->:1,$ s/<s>/!ENTER/g :1,$ s/<\/s>/!EXIT/g

Combine Acoustic and Language Model(2) HBuild 1.meaning:change the SRILM form to HTK network form 2.command: HBuild –n bigram_ascii.lm htkword.dic wd_net 3.htkword.dic is the relation between work and pronunciation EX:[B7C2] k_e e 4.wd_net is that word_net that we need EX:I=9996 W=[A64C][C3D2] J=570711 S=9551 E=2135 l=-7.52

Combine Acoustic and Language Model(3) HBuild(Cond.) - Questions htkword.dic has no word that corresponds to <unk> <sol>: don’t add flag –u when training language model, but is it appropriate?

Decode-HVite(1) 1.meaning:decode testing corpus 2.command: HVite –t 250 –H hmmset.mmf –S testing.scp –C hvite.cfg –w wd_net -i result.mlf –p 0.0 –s 0.0 htkword.dic rcd.lst 3.result is in result.mlf EX:0 1100000 !ENTER -1042.10318 1100000 2200000 [B37D] -734.694397

Decode-HVite(2) - Questions If we use HERest to train hmmset.mmf with phone.lst,it will error: ->cannot load file n_o <sol>:use rcd.lst instead

Divide Mixtures by HHEd But hmmset.mmf given by TA already has 32 mixtures for each state What config to use? Old DSP hw: the same as what HERest use My trail: it can run without config

Initailize Acoustic Model 2 methods for 2 different conditions With labels: use HInit and HRest Without labels: use HCompV Both of them has a similar question

Reference 1.HTK_Book.pdf 2.SpeechProject2007Winter.ppt 3.dsp_hw1.ppt(2006_spring_DSP) 4.HTK_Training.pdf 5.SpeechProj board in ptt2