EP0063454A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Method for recognizing machine encoded characters
Assignee: NCR Canada Ltd
•
Filing: 1982-04-08
•
Granted: —
•
Expires: 2002-04-08
Abstract
Abstract Machine encoded characters read from a document 10, such as a bank cheque, by an optical scanner 12 of a character recognition system 10 are stored in digitized form in a buffer 14. The stored digitized data is analyzed to find potential characters and to determine a reference location in respect of each potential character. The reference locations are stored and a first character recognition process is performed in order to determine if possible a machine encoded character for each reference location. The stored reference locations are analyzed in order to predict probable reference locations of additional potential characters, and a second recognition process is performed in order to determine if possible a machine encoded character for each predicted reference location. The method described reduces the number of characters that are rejected due to the presence of extraneous matter in the background of the characters being read.
Inventors
- 👤 John Stanley Grabowski
- 👤 Chin-Chih Shiau
- 👤 John Richard Latala
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-05 18:24
CPC Classifications
G; G06; G06V; G06V10/00; G06V10/20; G06V10/24; G06V10/245; G06V30/00; G06V30/10
Claims Analysis
Total Claims: 9
Claim 1
Independent Claim
128 words
A method for optically reading and determining machine encoded characters from a printed document in which an area of said printed document to be read is scanned to produce digitized data and said digitized data is stored, characterized by the steps of analyzing the stored digitized data in order to find potential characters and determining and storing a reference location in respect of each potential character found; performing a first character recognition process in order to determine if possible the machine encoded character for each such reference location; analyzing the stored reference locations in order to predict the probable reference location of any additional potential character; and performing a second character recognition process in order to determine if possible the machine encoded character for any predicted storage location.
Claim 2
Independent Claim
128 words
A method for optically reading and determining machine encoded characters from a printed document in which an area of said printed document to be read is scanned to produce digitized data and said digitized data is stored, characterized by the steps of analyzing the stored digitized data in order to find potential characters and determining and storing a reference location in respect of each potential character found; performing a first character recognition process in order to determine if possible the machine encoded character for each such reference location; analyzing the stored reference locations in order to predict the probable reference location of any additional potential character; and performing a second character recognition process in order to determine if possible the machine encoded character for any predicted storage location.
Claim 3
Dependent Claim
52 words
A method according to claim 1, characterized in that the step of analyzing said digitized data involves successively monitoring different portions of the stored digitized data in order to ascertain whether a predetermined set of conditions are satisfied, a reference location being stored each time said predetermined set of conditions are satisfied.
Claim 4
Dependent Claim
33 words
3 . A method according to claim 1, characterized in that the step of analyzing the stored reference locations serves to predict reference locations for blanks as well as any additional potential character.
Claim 5
Dependent Claim
161 words
4 . A method according to any one of the preceding claims, characterized in that each character recognition process includes the steps of: a) storing digitized data representing a potential character in a data storage matrix having a storage capacity greater than the storage capacity required to store a character; b) comparing digitized data stored in a first portion of said working matrix with one of a set of prestored masks which comprise digitized data representing character and non-character elements and which respectively represent possible characters, and determining the number of mismatched elements; c) storing the number of mismatched elements; d) repeating steps b) and c) for successive different portions of said data storage matrix until the total storage capacity of said data storage matrix has been covered; e) repeating steps b) to d) inclusive for different masks until all the prestored masks have been utilized; and f) determining the machine encoded character by analyzing the stored numbers of mismatched elements.
Claim 6
Dependent Claim
93 words
A method according to claim 4, characterized in that step b) includes the preliminary step of comparing the digitized data stored in said first portion of said data storage matrix with a prestored mask associated with said one of said set of prestored masks and indicative of non-significant elements of the stored data, in order to determine elements of the digitized data within said first portion of said data storage matrix which are of interest, only elements determined to be of interest being compared with said one of said set of prestored masks.
Claim 7
Dependent Claim
69 words
A method according to claim 4, characterized in that step f) involves determining the lowest number of mismatched elements found for each mask, determining which of these lowest numbers is the minimum number and the next minimum number, and determining whether said minimum number is less than a first preset value and whether the difference between said minimum number and said next minimum number exceeds a second preset value.
Claim 8
Dependent Claim
46 words
7 . A method according to claim 6, characterized in that each of said minimum number and said next minimum number has one of two possible values dependent on whether step f) forms part of said first character recognition process or said second character recognition process.
Claim 9
Dependent Claim
25 words
A method according to claim 1, characterized by the further step of providing an output in machine readable form representing each determined machine readable character.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.