Patent US4499596A successfully refreshed from Google Patents.
US4499596A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Adaptive facsimile compression using a dynamic extendable decision network
Assignee: International Business Machines Corp
•
Filing: 1982-06-28
•
Granted: 1985-02-12
•
Expires: 2002-06-28
Abstract
Abstract Preliminary classification of a text character by examining a few of its PELs rather than requiring repeated comparisons of the entire pattern with a large set of prototypes is achieved by following a path through a decision tree in a root-to-leaf node direction. Each tree node specifies the location of a PEL in the character and the input pattern PEL value in turn determines the next branch to follow in the tree. The root node identifies the first PEL to be inspected while the final leaf indicates the class selected. The library pattern and inputs are compared. Upon match, an index number is generated and the next pattern processed. Upon a mismatch, the pattern will be registered as a new prototype in the library, the tree extended to include the new class, and the next pattern will be processed.
Inventors
- 👤 Richard G. Casey
- 👤 Theodore D. Friedman
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-24 11:57
CPC Classifications
G; G06; G06F; G06F18/00; G06F18/20; G06F18/24; G06F18/243; G06F18/24323
Claims Analysis
Total Claims: 8
Claim 1
Independent Claim
179 words
A method for adaptively assigning index numbers to each one of a set of input PEL array patterns in a system formed from a scanner for elucidating input array patterns, a memory for storing pograms and data including a library of at least one pattern, output means, and a processor intercoupling the scanner, memory and output means, and responsive to the programs for operating upon the data, comprising the steps of: preclassifying an input pattern provided by the scanner according to one of the patterns in the library using color values in the input pattern at each location specified by a decision tree node in order to determine a successor node in a root-to-leaf node direction until terminating in an index number designating a library pattern; comparing the designated library pattern with the input and either upon a match, generating an index number and processing the next input pattern; or upon a mismatch, generaing a new index number, adding the input pattern to the library in the memory, and extending the decision tree; and processing the next input pattern.
Claim 2
Independent Claim
179 words
A method for adaptively assigning index numbers to each one of a set of input PEL array patterns in a system formed from a scanner for elucidating input array patterns, a memory for storing pograms and data including a library of at least one pattern, output means, and a processor intercoupling the scanner, memory and output means, and responsive to the programs for operating upon the data, comprising the steps of: preclassifying an input pattern provided by the scanner according to one of the patterns in the library using color values in the input pattern at each location specified by a decision tree node in order to determine a successor node in a root-to-leaf node direction until terminating in an index number designating a library pattern; comparing the designated library pattern with the input and either upon a match, generating an index number and processing the next input pattern; or upon a mismatch, generaing a new index number, adding the input pattern to the library in the memory, and extending the decision tree; and processing the next input pattern.
Claim 3
Independent Claim
234 words
A method for adaptively assigning index numbers to successive PEL into patterns in which the first pattern is used to initialize a pattern library and a decision tree, the tree leaf nodes being in concordance with index numbers, and in respect to second and subsequent input patterns in a system formed from a scanner for elucidating input array patterns, a memory for storing programs and data including a library of at least one pattern, output means, and a processor intercoupling the scanner, memory and output means, and responsive to the programs for operating upon the data, comprising the steps of: classifying a pattern aprovided by the scanner as one of the library patterns by reiteraively using the color value of each FEL whose input array location is spcified by a decision tree node to branch to a successor node in a root-to-leaf node direction until a leaf node is encountered; verfying the classification by comparison between the classified library pattern and the input pattern such that the PEL color value differences at the character edges are either within or without a predetermined correlation; and either generating an index number upon the comparison being within the correlation and processing the next input pattern; or generating a new index number, adding the pattern to the library in the memory, and extending the decision tree upon the comparison being without correlation and processing the next input pattern.
Claim 4
Independent Claim
234 words
A method for adaptively assigning index numbers to successive PEL into patterns in which the first pattern is used to initialize a pattern library and a decision tree, the tree leaf nodes being in concordance with index numbers, and in respect to second and subsequent input patterns in a system formed from a scanner for elucidating input array patterns, a memory for storing programs and data including a library of at least one pattern, output means, and a processor intercoupling the scanner, memory and output means, and responsive to the programs for operating upon the data, comprising the steps of: classifying a pattern aprovided by the scanner as one of the library patterns by reiteraively using the color value of each FEL whose input array location is spcified by a decision tree node to branch to a successor node in a root-to-leaf node direction until a leaf node is encountered; verfying the classification by comparison between the classified library pattern and the input pattern such that the PEL color value differences at the character edges are either within or without a predetermined correlation; and either generating an index number upon the comparison being within the correlation and processing the next input pattern; or generating a new index number, adding the pattern to the library in the memory, and extending the decision tree upon the comparison being without correlation and processing the next input pattern.
Claim 5
Dependent Claim
47 words
A method according to claim 1, wherein the step of extending the decision tree includes the step of identifying an insertion point, forming a node sub-network from reliable PEL array locations interior to the input pattern, and coupling the sub-network into the tree at the insertion point.
Claim 6
Dependent Claim
39 words
A method according to claim 2 wherein the insertion point is defined as the branch between a node specifying an unliable PEL array location and its antecedent node specifying the reliable PEL value both at the counterpart pattern locations.
Claim 7
Dependent Claim
233 words
A character recognition method implementable on an APL stored program controlled OCR processor responsive to a succession of character PEL arrays representing printed characters on a document after optical scanning, thresholding, and segmenting of their images, said OCR processor being formed from a scanner for elucidating character PEL arrays, a memory for storing programs and data including a library of at least one character array, output means, and a processing element intercoupling the scanner, memory and output responsive to the programs for operating upon the data, comprising the steps of: preclassifying an input character array provided by the scanner according to one of the character arrays in a library using color values in the input array at each location specified by a decision tree node in order to determine a successor node in a root-to-leaf node direction until terminating in an index number designating the library character; comparing the designated library character with the input character and either upon a match, generating an index number and processing the next input character; or upon a mismatch, generating a new index number, adding the new input character to the library in the memory, and extending the decision tree by identifying an insertion point, forming a node sub-network from reliable PEL array locations interior to the input character array, and coupling the sub-network into the tree at the insertion point; and processing the next input pattern.
Claim 8
Dependent Claim
233 words
A character recognition method implementable on an APL stored program controlled OCR processor responsive to a succession of character PEL arrays representing printed characters on a document after optical scanning, thresholding, and segmenting of their images, said OCR processor being formed from a scanner for elucidating character PEL arrays, a memory for storing programs and data including a library of at least one character array, output means, and a processing element intercoupling the scanner, memory and output responsive to the programs for operating upon the data, comprising the steps of: preclassifying an input character array provided by the scanner according to one of the character arrays in a library using color values in the input array at each location specified by a decision tree node in order to determine a successor node in a root-to-leaf node direction until terminating in an index number designating the library character; comparing the designated library character with the input character and either upon a match, generating an index number and processing the next input character; or upon a mismatch, generating a new index number, adding the new input character to the library in the memory, and extending the decision tree by identifying an insertion point, forming a node sub-network from reliable PEL array locations interior to the input character array, and coupling the sub-network into the tree at the insertion point; and processing the next input pattern.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1988-02-12 | 1988-08-13 | 1989-02-12 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1992-02-12 | 1992-08-13 | 1993-02-12 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 1996-02-12 | 1996-08-13 | 1997-02-12 | $7700.00 | Window Closed / Lapsed |