Non-edit multiple image font processing of records

• US 6,023,528 A
• Filed: 01/15/1999
• Issued: 02/08/2000
• Est. Priority Date: 10/28/1991
• Status: Expired due to Fees

• Current Assignee: Horst Froessl
• Original Assignee: Horst Froessl
• Inventors: Froessl, Horst
• Primary Examiner(s): Couso, Jose L.
• Application Number:
  US09/232,466
• Time in Patent Office: 389 Days
• Field of Search: 382/181, 382/229, 382/230, 382/231, 382/232, 382/309, 382/317, 382/135, 382/155, 382/276
• US Class Current: 382/181
• CPC Class Codes: G06F 16/40 of multimedia data, e.g. slideshows comprising image and additional audio data retrieval of still image data G06F16/50; retrieval of audio data G06F16/60; retrieval of video data G06F16/70
• CPC Class Codes: G06F 16/5846 using extracted text

Abstract

A sequence of documents is delivered to an optical scanner in which each document is scanned to form a digital image representation of the content of the document. The image representation is automatically examined by data processing apparatus to select search words which meet predetermined criteria and by which the document can be subsequently located. The search words are stored in a non-volatile memory and the entire document content is stored in mass storage in image form. A font table is established and images of entered search words are constructed from the table. Unrecognized or imperfectly formed ambiguous characters are stored with the font table and are used in the construction of search words to eliminate text editing or are stored with converted characters of search words. The approach can also be used to eliminate editing of full text-converted data bases during input.

Claims

1. A method of inputting and preparing data from source documents for storage and subsequent retrieval comprising the steps of

scanning each source document and forming signals representative of digitized patterns derived from images of characters and graphics thereon,

storing the signals representative of the digitized patterns,

selecting segments of the stored signals for further processing,

converting signals representative of digitized patterns of characters into a machine code,

storing as ambiguous characters the digitized patterns of each character not successfully converted into machine code including storing as ambiguous words each group of characters which includes converted characters and at least one ambiguous character, and

storing the digitized patterns of the selected segments correlated with the machine code for subsequent use.

2. A method according to claim 1 and including

displaying groups of characters which include converted characters and at least one ambiguous character, and

manually identifying the ambiguous character.

3. A method according to claim 1 and further including preparing for a search by

displaying a selected group of converted characters to be used as a search word,

concurrently displaying the image of the group from which the characters were converted, and

conducting the search with the image.

4. A method according to claim 1 and including, if the selected group includes an ambiguous character which is recognizable on the display, manually correcting the ambiguous character by keying in the correct character.

5. A method according to claim 4 and including, if the selected group includes an ambiguous character and is the same word as the keyed-in word, approving the displayed word and thereby correcting the converted group.

6. A method according to claim 4 and including, if the selected group includes an ambiguous character which is not recognizable on the display, additionally displaying words adjacent to the displayed word from the image of the document from which the word with the ambiguous character was selected.

7. A method according to claim 4 and including displaying an ambiguous character in a distinctive display from the converted characters.

8. A method according to claim 4 and including displaying the group of characters with recognized characters displayed from code and ambiguous characters displayed in image.

9. A method according to claim 3 and including

automatically machine-selecting search words from each document to be used in locating the document from mass storage.

10. A method according to claim 9 and including correlating each search word with a date indicative of an interval of time during which the document from which the search word was selected was sent and received.

11. A method according to claim 9 and further comprising

forming a logo table of stored patterns of images of logo designs identifying sending organizations together with information in code form about the sender employing each such design,

when a document having a design is scanned, conducting a pattern search of the stored patterns of images in the logo table to seek a match between the scanned design and a stored image,

when a pattern match is found, retrieving and correlating with the identification of the document the identifying organization information associated with the matched pattern from the logo table, and

when a match is not found, flagging the document for addition of the design and identifying company information to the logo table.

12. A method according to claim 3 which includes selecting patterns of images of ambiguous characters for use as search word images.

13. A method according to claim 12 and including combining recognized characters with ambiguous characters to form search words.

14. A method according to claim 13 and including selecting a search word by keying in a word, displaying a stored image of a selected group of characters of a similar word and, if the selected group includes an ambiguous character and is the same word as the keyed in word, approving the displayed word and thereby correcting the ambiguous character in the converted group.

15. A method according to claim 3 wherein the step of storing patterns of images of ambiguous characters includes storing the pattern of the image of each group of characters containing an ambiguous character.

16. A method according to claim 3 and further comprising

establishing a font table in memory including patterns of images of characters in a plurality of different fonts, each character of each font having a unique, identifiable location in a memory area, and

comparing characters of an incoming document with the patterns of images in the font table and, if a matching font is not found, adding patterns of images of the characters of the incoming document to the font table.

17. A method according to claim 16 and including determining an identity for the sender of each document scanned, and correlating the identity with specific fonts used by the sender.

18. A method according to claim 16 and further including establishing a dictionary of search words in image for each font in the font table, and wherein the step of selectively recognizing and converting groups of characters includes comparing groups of characters in each scanned document with the dictionary to determine which words should be converted into code, and storing converted words as search words.

19. A method according to claim 18 wherein the dictionary includes words which are to be selected as search words.

20. A method according to claim 18 wherein the dictionary includes words which are not to be selected as search words.