Automated search on cursive records not having an ASCII index

• US 6,917,709 B2
• Filed: 12/16/2002
• Issued: 07/12/2005
• Est. Priority Date: 07/20/1999
• Status: Expired due to Term

• Current Assignee: Parascript, LLC
• Original Assignee: Parascript, LLC
• Inventors: Zelinski, Pavel
• Primary Examiner(s): Ahmed, Samir
• Assistant Examiner(s): BHATNAGAR, ANAND P
• Application Number:
  US10/321,099
• Publication Number:
  US 20030099399A1
• Time in Patent Office: 939 Days
• Field of Search: 382119-123, 382155-161, 382/179, 382/186, 382/217, 382/218, 382/305, 382/306, 707/3, 707/6
• US Class Current: 382/186
• CPC Class Codes: G06F 16/5846 using extracted text
• CPC Class Codes: G06V 30/2268 using stroke segmentation
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

Images of handwritten cursive records are extracted, and an automated search on the images of the cursive records is performed based on an ASCII query for a record. A cursive equivalent of the ASCII query is matched to the images of the cursive records, and a similarity value is generated to indicate the extent of match between features of the cursive equivalent of the ASCII query and features of each cursive record. The records are sorted based upon their similarity value determined in the matching process. This provides a candidate list of cursive record images to be manually examined by a user for the purpose of making a final determination as to whether any of the cursive records on the candidate list satisfy the query.

Claims

1. A method for finding a desired handwritten computer record using automated searching by a computing system comprising:

extracting a snippet of each handwritten computer record and identifying each snippet with an associated handwritten computer record;

automated searching of the snippets by the computing system to select best matches between a query and snippets extracted by the extracting act, the best matches forming a candidate list; and

displaying the candidate list of snippets that match the query close enough to warrant display of the associated handwritten computer records whereby the number of handwritten computer records, that must be displayed to find the desired handwritten computer record, are reduced.

2. The method of claim 1 wherein the act of extracting comprises:

identifying a source handwritten computer record;

extracting the snippet out of the source handwritten computer record; and

storing the snippet and an identification of its source handwritten computer record in snippet description data.

3. The method of claim 2 wherein the act of automated searching comprises:

matching a query against each snippet and generating a similarity value indicative of how well the query matches the snippet;

sorting query matches by similarity value; and

selecting query matches with the similarity values above a certain value to create a candidate list of best matches.

4. The method of claim 3 wherein the act of displaying comprises:

looking up snippets on the candidate list; and

retrieving source handwritten computer records for snippets on the candidate list that best match the query.

5. A computer search method for performing an automated search of handwritten computer records, the search method comprising:

extracting handwritten record phrases from a handwritten computer record;

converting a query into a plurality of handwritten variations for characters in the query to generate a plurality of handwritten search features associated with the query;

matching the handwritten search features against features of the handwritten record phrases in the handwritten computer record and generating a similarity value for each handwritten record phrase as a measure of similarity between the features of the handwritten record phrase and the handwritten search features associated with the query;

selecting handwritten record phrases based on their similarity values to create a candidate list of handwritten record phrases as possible candidate matches for the query; and

displaying the candidate list of handwritten record phrases.

6. The search method of claim 5 wherein handwritten search features are a sequence of handwritten character stroke features and each sequence is a variation from other sequences of handwritten character stroke features for the query due to variations in handwriting styles and phonetic variations in the query.

7. The search method of claim 5 wherein the act of selecting handwritten record phrases comprises:

sorting the handwritten record phrases into an ordered list according to similarity value;

setting a threshold value;

selecting for the candidate list the handwritten record phrases having a similarity value that satisfies the threshold value.

8. The search method of claim 5 wherein the act of selecting handwritten record phrases comprises:

sorting the handwritten record phrases into an ordered list according to similarity value from most similar record phrase to least similar record phrase;

selecting a predetermined number of the most similar handwritten record phrases from the ordered list for the candidate list.