KR970050114A - Keyword detection method using similarity between keywords - Google Patents

Keyword detection method using similarity between keywords Download PDF

Info

Publication number
KR970050114A
KR970050114A KR1019950053945A KR19950053945A KR970050114A KR 970050114 A KR970050114 A KR 970050114A KR 1019950053945 A KR1019950053945 A KR 1019950053945A KR 19950053945 A KR19950053945 A KR 19950053945A KR 970050114 A KR970050114 A KR 970050114A
Authority
KR
South Korea
Prior art keywords
similarity
key
key words
key word
words
Prior art date
Application number
KR1019950053945A
Other languages
Korean (ko)
Other versions
KR100211067B1 (en
Inventor
안영목
Original Assignee
양승택
한국전자통신연구원
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 양승택, 한국전자통신연구원 filed Critical 양승택
Priority to KR1019950053945A priority Critical patent/KR100211067B1/en
Publication of KR970050114A publication Critical patent/KR970050114A/en
Application granted granted Critical
Publication of KR100211067B1 publication Critical patent/KR100211067B1/en

Links

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G10L2015/088—Word spotting

Landscapes

  • Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Telephonic Communication Services (AREA)

Abstract

본 발명은 핵심어 검출 음성 인식 시스템에서 핵심어간의 유사도를 이용하여 핵심어를 검출하는 핵심어 검출 방법에 관한 것으로, 핵심어 승인 판단부에서 등록 단어간의 유사도를 참고하여 현재 인식된단어를 출력할 것인지 혹은 거절할 것인지 판단하여 그 승인 여부를 인식 결과 출력부에 전달하는 핵심어 검출 방법을 제공하기 위하여, 핵심어 발생 확률값을 구하여 최고 발생 확률값을 갖는 핵심어를 검출한 후에 타 핵심어와의 유사도를 구하고, 핵심어간의 유사도 가능값을 입력받아 상기 두 유사도를 비교하여 검출된 핵심어간의 유사도가 발생 가능한가를 판단하는 제1단계(31 내지 35); 핵심어간의 유사도 발생이 가능하지 않으면 출력을 거절하는 제2단계(36); 및 핵심어간의 유사도 발생이 가능하면 상기 제1단계(31 내지 35)의 유사도 판단 과정(35)을 등록되어 잇는 모든 핵심어에 대해서 반복한 후에 출력을 승인하는 제3단계(37,38)를 포함하여 핵심어 검출 및 비핵심어에 대한 거절이 보다 유연성을 갖도록 하는 효과가 있다.The present invention relates to a key word detection method for detecting a key word using similarity between key words in a key word detection speech recognition system. The key word recognition determination unit outputs or rejects a currently recognized word with reference to the similarity between registered words. In order to provide a method for detecting a key word that determines and transmits the recognition result to a recognition result output unit, a key word occurrence probability value is obtained, the key word having the highest occurrence probability value is detected, the similarity with other key words is obtained, and the similarity possible values between the key words are obtained. A first step (31 to 35) of receiving an input and comparing the two similarities to determine whether similarity between the detected key words can occur; A second step 36 of rejecting output if similarity between key words is not possible; And a third step (37, 38) of approving the output after repeating the similarity determination process (35) of the first step (31 to 35) for all key words registered if the similarity between key words is possible. Key word detection and rejection of non-key words have the effect of making them more flexible.

Description

핵심어간의 유사도를 이용한 핵심어 검출 방법Keyword detection method using similarity between keywords

본 내용은 요부공개 건이므로 전문내용을 수록하지 않았음Since this is an open matter, no full text was included.

제2도는 본 발명이 적용되는 핵심어 검출 음성인식 시스템의 구성도.2 is a block diagram of a keyword detection speech recognition system to which the present invention is applied.

제3도는 본 발명에 따른 흐름도.3 is a flow chart according to the present invention.

Claims (1)

핵심어 검출 음성 인식 시스템에 적용되는 핵심어 검출 방법에 있어서, 핵심어 발생 확률값을 구하여 최고 발생 확률값을 갖는 핵심어를 검출한 후에 타 핵심어와의 유사도를 구하고, 핵심어간의 유사도 가능값을 입력받아 상기 두 유사도를 비교하여 검출된 핵심어간의 유사도가 발생 가능한가를 판단하는 제1단계(31 내지 35); 상기 제1단계(31 내지 35)의 판단 결과, 핵심어간의 유사도 발생이 가능하지 않으면 출력을 거절하는 제2단계(36); 및 상기 제1단계(31 내지 35)의 판단 결과, 핵심어간의 유사도 발생이 가능하면 상기 제1단계(31 내지 35)의 유사도 판단 과정(35)을 등록되어 있는 모든 핵심어에 대해서 반복한 후에 출력을 승인하는 제3단계(37,38)를 포함하는 것을 특징으로 하는 핵심어 검출 방법.Key word detection In a key word detection method applied to a speech recognition system, a key word occurrence probability value is detected, a key word having the highest occurrence probability value is detected, a similarity with other key words is obtained, and similarity values between the key words are received and the two similarities are compared. First steps 31 to 35 to determine whether similarity between the detected key words can occur; A second step (36) of rejecting output if similarity between key words is not possible as a result of the determination of the first step (31 to 35); As a result of the determination of the first steps 31 to 35, if similarity between key words is possible, the similarity determination process 35 of the first steps 31 to 35 is repeated for all registered key words, and then output is output. And a third step (37, 38) of approving. ※ 참고사항 : 최초출원 내용에 의하여 공개하는 것임.※ Note: The disclosure is based on the initial application.
KR1019950053945A 1995-12-22 1995-12-22 Keyword detecting method using proximity between keywords KR100211067B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1019950053945A KR100211067B1 (en) 1995-12-22 1995-12-22 Keyword detecting method using proximity between keywords

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1019950053945A KR100211067B1 (en) 1995-12-22 1995-12-22 Keyword detecting method using proximity between keywords

Publications (2)

Publication Number Publication Date
KR970050114A true KR970050114A (en) 1997-07-29
KR100211067B1 KR100211067B1 (en) 1999-07-15

Family

ID=19442774

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1019950053945A KR100211067B1 (en) 1995-12-22 1995-12-22 Keyword detecting method using proximity between keywords

Country Status (1)

Country Link
KR (1) KR100211067B1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR970071462A (en) * 1996-04-15 1997-11-07 김광호 Method for rejecting recognition result in speech recognition system
KR100482313B1 (en) * 1996-12-30 2005-07-21 엘지전자 주식회사 Speech Recognition Method Using Dual Similarity Comparison
US8032374B2 (en) 2006-12-05 2011-10-04 Electronics And Telecommunications Research Institute Method and apparatus for recognizing continuous speech using search space restriction based on phoneme recognition

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR970071462A (en) * 1996-04-15 1997-11-07 김광호 Method for rejecting recognition result in speech recognition system
KR100482313B1 (en) * 1996-12-30 2005-07-21 엘지전자 주식회사 Speech Recognition Method Using Dual Similarity Comparison
US8032374B2 (en) 2006-12-05 2011-10-04 Electronics And Telecommunications Research Institute Method and apparatus for recognizing continuous speech using search space restriction based on phoneme recognition

Also Published As

Publication number Publication date
KR100211067B1 (en) 1999-07-15

Similar Documents

Publication Publication Date Title
Wacholder et al. Disambiguation of proper names in text
KR900018909A (en) Language recognition method and language recognizer training method
JPS62217295A (en) Voice recognition system
KR970050114A (en) Keyword detection method using similarity between keywords
Xu et al. Keyword search using query expansion for graph-based rescoring of hypothesized detections
KR910017312A (en) Information retrieval system using approximate match between input string and keyword and matching method
JP2802602B2 (en) Dialogue processing device
Islam et al. A generalized approach to word segmentation using maximum length descending frequency and entropy rate
JPH0256086A (en) Method for postprocessing for character recognition
KR19990015114A (en) Character Recognizer Using Character Connection Information
JPH0520354A (en) Japanese-language sentence analyzer
JPH01156800A (en) Voice recognition apparatus
Ajees et al. A hybrid approach for suffix separation in malayalam
JPH03116376A (en) Keyword matching device
JP2923295B2 (en) Pattern identification processing method
JPH0344316B2 (en)
JPH0792675B2 (en) Voice recognizer
JP3446769B2 (en) Character recognition device
JP2005242181A (en) Method and apparatus for vocabulary selection, and program
JPH10340096A (en) Voice recognition device
JPS63173100A (en) Keyword extractor
JPH01253799A (en) Recognizing method for speech
JPH10171799A (en) Method and device for analyzing name
Jiménez et al. Instance metrics improvement by probabilistic support
KR970019272A (en) Name Search Method in Voice Dialing System

Legal Events

Date Code Title Description
A201 Request for examination
E902 Notification of reason for refusal
E701 Decision to grant or registration of patent right
GRNT Written decision to grant
FPAY Annual fee payment

Payment date: 20080328

Year of fee payment: 10

LAPS Lapse due to unpaid annual fee