Work Files Saved Searches
   My Account                                                  Search:   Quick/Number   Boolean   Advanced       Help   

 The Delphion Integrated View

  Buy Now:   Buy PDF- 17pp  PDF  |   File History  |   Other choices   
  Tools:  Citation Link  |  Add to Work File:    
  View:  Expand Details   |  INPADOC   |  Jump to: 
 Email this to a friend  Email this to a friend 
Title: US5502774: Automatic recognition of a consistent message using multiple complimentary sources of information
[ Derwent Title ]

Country: US United States of America

View Images High


17 pages

Inventor: Bellegarda, Jerome R.; Goldens Bridge, NY
Kanevsky, Dimitri; Ossining, NY

Assignee: International Business Machines Corporation, Armonk, NY
other patents from INTERNATIONAL BUSINESS MACHINES CORPORATION (280070) (approx. 44,393)
 News, Profiles, Stocks and More about this company

Published / Filed: 1996-03-26 / 1994-09-06

Application Number: US1994000300232

IPC Code: Advanced: G06K 9/62; G06K 9/68; G10L 15/24; G10L 15/06; G10L 15/10;
IPC-7: G06K 9/00;

ECLA Code: G06K9/62F3M; G10L15/24; S10L15/063M; S10L15/10;

U.S. Class: Current: 382/159; 382/116; 382/187; 382/228; 704/231; 704/232; 704/235; 704/240; 704/241; 704/256; 704/256.3; 704/E15.041;
Original: 382/159; 382/116; 382/187; 382/228; 381/043;

Field of Search: 381/001,2,3,13,14,15,30,36,38,7,39,116,157,159,187,227,228 381/041,43 340/825.3

Priority Number:
1994-09-06  US1994000300232
1992-06-09  US1992000895967

Abstract:     A general approach is provided for the combined use of several sources of information in the automatic recognition of a consistent message. For each message unit (e.g., word) the total likelihood score is assumed to be the weighted sum of the likelihood scores resulting from the separate evaluation of each information source. Emphasis is placed on the estimation of weighing factors used in forming this total likelihood. This method can be applied, for example, to the decoding of a consistent message using both handwriting and speech recognition. The present invention includes three procedures which provide the optimal weighing coefficients.

Attorney, Agent or Firm: Sterne, Kessler, Goldstein & Fox ;

Primary / Asst. Examiners: Boudreau, Leo; Anderson, David R.

INPADOC Legal Status: Show legal status actions          Buy Now: Family Legal Status Report

Related Applications: Go to Result Set: 1 patent(s) that list this one as related
Application Number Filed Patent Pub. Date  Title
US1992000895967 1992-06-09       

Parent Case:     This application is a continuation, of application Ser. No. 07/895,967, filed Jun. 9, 1992, now abandoned.

Family: Show 2 known family members

First Claim:
Show all 28 claims
Having thus described our invention, what we claim as new and desire to secure by Letters Patent is:     1. In a message recognition system, a method of transforming a consistent message into a message recognizable by the computer, the method comprising the steps of:
  • (A) transforming the consistent message generated by a human in at least two formats into electrical signal representations of the consistent message;
  • (B) producing from said electrical signal representations of the consistent message a set of parameters for each said format;
  • (C) generating a likelihood score of recognition for each said set of parameters;
  • (D) using said sets of parameters to train a weighting coefficient for each of the at least two formats of the consistent message, wherein said step of training a weighting coefficient comprises the steps of
    • (i) partitioning the consistent message in each said format into one or more subunits, wherein each subunit corresponds to a piece of the consistent message,
    • (ii) grouping said subunits from each said format into a plurality of groups, wherein each group comprises one said subunit from each said format, and wherein each said subunit in one said group corresponds to the same piece of the consistent message,
    • (iii) determining a likelihood score of recognition for each said group of subunits,
    • (iv) determining a global score for the consistent message based on said likelihood scores of recognition, and
    • (v) using said global score to determine said weighting coefficients,
  • (E) generating a weighted expression based on said trained weighting coefficient and said likelihood scores of recognition; and
  • (F) selecting a candidate message unit that maximizes said weighted expression to transform said electrical signal representations of the consistent message into a computer recognizable message.

Background / Summary: Show background / summary

Drawing Descriptions: Show drawing descriptions

Description: Show description

Forward References: Show 113 U.S. patent(s) that reference this one

U.S. References: Go to Result Set: All U.S. references   |  Forward references (113)   |   Backward references (12)   |   Citation Link

Patent  Pub.Date  Inventor Assignee   Title
Get PDF - 10pp USRE31692  1984-10 Tyhurski et al.  Optical Recognition Systems, Inc. Combined magnetic optical character reader
Get PDF - 6pp US3192321  1965-06 Nassimbene   Electronic lip reader
Get PDF - 15pp US4348553  1982-09 Baker et al.  International Business Machines Corporation Parallel pattern verifier with dynamic time warping
Get PDF - 27pp US4481593  1984-11 Bahler  Exxon Corporation Continuous speech recognition
Get PDF - 12pp US4731857  1988-03 Tappert  International Business Machines Corporation Recognition system for run-on handwritten characters
Get PDF - 22pp US4754489  1988-06 Bokser  The Palantir Corporation Means for resolving ambiguities in text based upon character context
Get PDF - 14pp US4829577  1989-05 Kuroda et al.  International Business Machines Corporation Speech recognition method
Get PDF - 9pp US4876731  1989-10 Loris et al.  Nynex Corporation Neural network model in pattern recognition using probabilistic contextual information
Get PDF - 22pp US4975960  1990-12 Petajan   Electronic facial tracking and detection system and method and apparatus for automated speech recognition
Get PDF - 13pp US4993068  1991-02 Piosenka et al.  Motorola, Inc. Unforgeable personal identification system
Get PDF - 13pp US5159644  1992-10 Martin et al.  Eastman Kodak Company Character recognition method using statistically weighted correlation
Get PDF - 10pp US5212821  1993-05 Gorin et al.  AT&T Bell Laboratories Machine-based learning system
Foreign References:
Publication Date IPC Code Assignee   Title
Buy PDF- 9pp GB2229305 1990-09  A61B 5/117 BRITISH TELECOMM Identity verification 

Other Abstract Info: DERABS G1996-179562 DERABS G1996-179562 DERABS G1997-235477

Inquire Regarding Licensing

Powered by Verity

Plaques from Patent Awards      Gallery of Obscure PatentsNominate this for the Gallery...

Thomson Reuters Copyright © 1997-2014 Thomson Reuters 
Subscriptions  |  Web Seminars  |  Privacy  |  Terms & Conditions  |  Site Map  |  Contact Us  |  Help