Method and apparatus for context-aided recognition

• US 4,058,795 A
• Filed: 12/23/1974
• Issued: 11/15/1977
• Est. Priority Date: 10/03/1972
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Balm, Gerald John
• Primary Examiner(s): Boudreau, Leo H.
• Application Number:
  US05/535,401
• Time in Patent Office: 1,058 Days
• Field of Search: 340/146.3 MA, 340/146.3 AQ, 340/146.3 Q, 340/146.3 WD, 340/146.3 ED, 340/146.3 R, 340/172.5
• US Class Current: 382/230
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/262 using context analysis, e.g. lexical, syntactic or semantic context

Abstract

A character-recognition unit or other input device provides a sequence of input characters. A questionable identification of one of the input characters is resolved by considering surrounding characters by a modified N-gram technique. The context characters are each translated into one of a small number of reduced codes each representing a subset of the alphabet to be recognized. The reduced codes are combined to produce a context code which indicates a set of categories containing the correct identification for the questionable input character. A list of confusion codes specifies a number of possibly correct identifications for one of the input characters. One of the codes in the confusion list is selected by the context code as the correct identification for the questionable character.

Claims

1. A system for enhancing the recognition of characters, comprising:

recognition means responsive to a sequence of character patterns for producing a sequence of character codes identifying said characters with respect to a predetermined plurality of categories;

coding means coupled to said recognition means for translating a plurality of said character codes to a plurality of reduced codes corresponding to a plurality of sets of said categories, the number of said sets being substantially smaller than the number of said categories;

first storage means responsive to said reduced codes for producing a context code representing one set of said categories;

second storage means for holding a plurality of confusion codes associated with a predetermined one of said character codes, each of said confusion codes identifying one of said categories; and

gating means responsive to said context code for selecting one of said confusion codes.

2. The system of claim 1 wherein said recognition means is further adapted to produce a marker flag associated with a non-confident one of said character codes;

and wherein said gating means is further responsive to the absence of said marker flag for inhibiting said one confusion code.

3. The system of claim 1, wherein said recognition means is further adapted to produce marker flags associated with non-confident ones of said character codes;

and wherein said gating means is further responsive to a predetermined combination of said marker flags for selecting a reject code and inhibiting said one confusion code.

4. The system of claim 1, wherein said first storage means comprises a storage array for producing a plurality of provisional codes, and logic means for selecting one of said provisional codes as said context code.

5. The system of claim 4, wherein said gating means is coupled to said logic means for substituting a reject code for said context code when said provisional codes bear a predetermined relationship with one another.

6. The system of claim 1, further comprising feedback means for substituting said one confusion code for a predetermined one of said character code in said sequence.

7. The system of claim 1, further comprising:

third storage means responsive to a plurality of said character codes for producing flags for predetermined combinations of said character codes; and

means responsive to said flags for modifying one of said character codes.

8. The system of claim 1, further comprising:

dictionary means responsive to a plurality of said character codes for producing a further character code;

detecting means responsive to a predetermined combination of said character codes for producing a small-word flag; and

means coupled to said detecting means for substituting said further character code for said one confusion code.