EP0555023A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Pattern recognition
Assignee: Canon Inc
•
Filing: 1993-01-29
•
Granted: —
•
Expires: 2013-01-29
Abstract
Abstract Pattern recognition apparatus and method which provides plural pattern recognition modes, one of which is selected in accordance with characteristics of an unknown pattern to be recognized. The pattern recognition apparatus includes a pattern recognizer having plural recognition modes for comparing an unknown pattern with at least one known pattern, and a selector for selecting one of the plural recognition modes based on characteristics of the unknown pattern. In one aspect, the selection of one of the plural recognition modes is based on font size, font type and image quality, and in another aspect, selection of the recognition mode is based on the results of prior pattern recognition operations. The pattern recognition apparatus may be provided in a stand-alone device, and this device may interface to a local area network bus. Alternatively, the pattern recognition apparatus may be distributed such that individual components thereof are provided independently on a local area network bus and provided with individual servers for interface with stand-alone computers. In the case where computers perform character processing or word processing, a preview mode may be provided so that the results of character recognition may be previewed before transmitting image information to a remote location.
Inventors
- 👤 Toshiaki c/o Canon Inf. Systems Inc. Yagasaki
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-09 10:35
CPC Classifications
G; G06; G06V; G06V30/00; G06V30/10; G06V30/24; G06V30/242
Claims Analysis
Total Claims: 52
Claim 1
Independent Claim
46 words
Apparatus for recognizing unknown patterns comprising: a pattern recognizer for comparing the unknown pattern with at least one known pattern, the pattern recognizer having plural recognition modes; and a selector for selecting one of the plural recognition modes based on the characteristics of the unknown pattern.
Claim 2
Dependent Claim
37 words
Apparatus according to Claim 1, further comprising feature extraction means for extracting a feature vector from the unknown pattern, wherein the pattern recognizer compares feature the vectors of the unknown pattern with feature vectors of known patterns.
Claim 3
Dependent Claim
24 words
Apparatus according to Claim 1, wherein said selector selects one of the plural recognition modes based on the results of prior pattern recognition operations.
Claim 4
Dependent Claim
38 words
Apparatus according to Claim 1, wherein said pattern recognizer includes first and second dictionaries, and wherein said plural recognition modes include a first mode which utilizes said first dictionary and a second mode which utilizes said second dictionary.
Claim 5
Dependent Claim
47 words
Apparatus according to Claim 4, further comprising reject counting means wherein said selector is adapted initially to select said first mode and, in a case where said reject counting means counts greater than a predetermined number of rejects, said selector is adapted to select said second mode.
Claim 6
Dependent Claim
23 words
Apparatus according to Claim 5, wherein said first dictionary corresponds to good quality patterns, and wherein said second dictionary corresponds to degraded patterns.
Claim 7
Dependent Claim
20 words
Apparatus according to Claim 4, wherein one of the first and the second modes is selected based on pattern quality.
Claim 8
Dependent Claim
20 words
Apparatus according to Claim 1, wherein each of said plural recognition modes uses a different discriminant function for recognizing patterns.
Claim 9
Dependent Claim
24 words
Apparatus according to Claim 8, wherein one of said different discriminant functions is selected in accordance with font size, font type and image quality.
Claim 10
Dependent Claim
46 words
Apparatus for recognizing an unknown pattern comprising: a pattern recognizer having plural different discriminant functions by which a candidate is assigned to the unknown pattern; and a selector for selecting one of said plural different discriminant functions in accordance with a characteristic of the unknown pattern.
Claim 11
Dependent Claim
29 words
Apparatus according to Claim 10, wherein said selector is adapted to select one of said plural different discriminant functions in accordance with font size, font type, and image quality.
Claim 12
Dependent Claim
26 words
Apparatus according to Claim 11, wherein said selector selects a Mahalanobis discriminant function for a degraded pattern and for a pattern having a small type face.
Claim 13
Dependent Claim
27 words
Apparatus according to Claim 11, wherein said selector selects a mean absolute distance discriminant function for a clear pattern and for a pattern having a large typeface.
Claim 14
Dependent Claim
31 words
Apparatus according to Claim 10, further comprising reject counting means, wherein one of said plural different discriminant functions is selected in accordance with the number counted by said reject counting means.
Claim 15
Dependent Claim
62 words
Apparatus for recognizing an unknown pattern comprising: a pattern recognizer including first and second dictionaries each for storing at least one known pattern, said pattern recognizer for comparing the unknown pattern with the known pattern in one of said first and second dictionaries; and a selector for selecting one of the first and second dictionaries based on characteristics of the unknown pattern.
Claim 16
Dependent Claim
41 words
Apparatus according to Claim 15, wherein said selector selects the first dictionary in the case where the unknown pattern is a good quality pattern and wherein said selector selects said second dictionary in the case where said unknown pattern is degraded.
Claim 17
Dependent Claim
47 words
Apparatus according to Claim 15, further comprising reject counting means wherein said selector is adapted initially to select said first dictionary and, in the case where said reject counting means counts greater than a predetermined number of rejects, said selector is adapted to select said second dictionary.
Claim 18
Dependent Claim
62 words
Pattern recognition apparatus comprising: a host for inputting image information and for segmenting image information into pattern serpents; and pattern recognition means for providing a candidate for a pattern segment transmitted from said host, said pattern recognition means including plural recognition modes, one of said plural recognition modes being selected in accordance with characteristics of the pattern segment transmitted from said host.
Claim 19
Dependent Claim
31 words
Apparatus according to Claim 18, further comprising a network bus to which said host is connected, and wherein said host is responsive to commands on said network bus for pattern recognition.
Claim 20
Dependent Claim
51 words
Apparatus according to Claim 19, further comprising reject counting means, wherein said pattern recognition means is adapted initially to select a first mode and, in a case where said reject counting means counts greater than a predetermined number of rejects, said pattern recognition means is adapted to select said second mode.
Claim 21
Dependent Claim
43 words
Apparatus according to Claim 18, further comprising discrimination means for discriminating between character patterns and non-character patterns in said image information, and wherein said pattern recognition means provides a candidate in the case where a character pattern is discriminated by said discriminating means.
Claim 22
Dependent Claim
37 words
Apparatus according to Claim 18, further comprising information processing means for processing character information, and further comprising preview means for causing character information processed by said information processing means to be processed by said pattern recognition means.
Claim 23
Dependent Claim
113 words
Pattern recognition apparatus comprising: a local area network bus; an image input device connected to said network bus, said image input device for inputting image information; pattern recognition means connected to said network bus, said pattern recognition means for providing a candidate for an unknown pattern and including plural recognition modes, one of said plural recognition modes being selected in accordance with characteristics of the unknown pattern; and plural processing means connected to said network bus, each said processing means for segmenting image information input by said image input device into segmented patterns, for providing a segmented pattern to said pattern recognition means, and for receiving the candidate from said pattern recognition means.
Claim 24
Dependent Claim
51 words
Apparatus according to Claim 23, further comprising reject counting means, wherein said pattern recognition means is adapted initially to select a first mode and, in a case where said reject counting means counts greater than a predetermined number of rejects, said pattern recognition means is adapted to select a second mode.
Claim 25
Dependent Claim
43 words
Apparatus according to Claim 23, further comprising discrimination means for discriminating between character patterns and non-character patterns in said image information, and wherein said pattern recognition means provides a candidate in the case where a character pattern is discriminated by said discriminating means.
Claim 26
Dependent Claim
43 words
Apparatus according to Claim 23, wherein each of said plural processing means includes information processing means for processing character information, and further comprising preview means for causing character information processed by said information processing means to be processed by said pattern recognition means.
Claim 27
Independent Claim
41 words
A method for recognizing unknown patterns comprising the steps of: selecting one of plural recognition modes based on the characteristics of the unknown pattern; and comparing the unknown pattern with at least one known pattern in the selected plural recognition mode.
Claim 28
Dependent Claim
34 words
A method according to Claim 27, further comprising the step of extracting a feature vector from the unknown pattern, and comparing the feature vector of the unknown pattern with feature vectors of known patterns.
Claim 29
Dependent Claim
26 words
A method according to Claim 27, wherein said selecting step selects one of the plural recognition modes based on the results of prior pattern recognition operations.
Claim 30
Dependent Claim
32 words
A method according to Claim 27, wherein said selecting step selects from among at least a first mode which utilizes a first dictionary and a second mode which utilizes a second dictionary.
Claim 31
Dependent Claim
55 words
A method according to Claim 30, further comprising the step of counting a number of rejects, wherein said selecting step is adapted initially to select said first mode and, in a case where said reject counting step counts greater than a predetermined number of rejects, said selecting step is adapted to select said second mode.
Claim 32
Dependent Claim
24 words
A method according to Claim 31, wherein said first dictionary corresponds to good quality patterns, and wherein said second dictionary corresponds to degraded patterns.
Claim 33
Dependent Claim
21 words
A method according to Claim 30, wherein one of the first and the second modes is selected based on pattern quality.
Claim 34
Dependent Claim
21 words
A method according to Claim 27, wherein each of said plural recognition modes uses a different discriminant function for recognizing patterns.
Claim 35
Dependent Claim
29 words
A method according to Claim 34, wherein in said selecting step one of said different discriminant functions is selected in accordance with font size, font type and image quality.
Claim 36
Independent Claim
40 words
A method for recognizing an unknown pattern comprising the steps of: selecting one of plural different discriminant functions in accordance with a characteristic of the unknown pattern; and assigning a candidate to the unknown pattern using the selected discriminant function.
Claim 37
Dependent Claim
28 words
A method according to Claim 36, wherein said selecting step selects one of said plural different discriminant functions in accordance with font size, font type, and image quality.
Claim 38
Dependent Claim
28 words
A method according to Claim 37, wherein said selecting step selects a Mahalanobis discriminant function for a degraded pattern and for a pattern having a small type face.
Claim 39
Dependent Claim
29 words
A method according to Claim 37, wherein said selecting step selects a mean absolute distance discriminant function for a clear pattern and for a pattern having a large typeface.
Claim 40
Dependent Claim
37 words
A method according to Claim 36, further comprising the step of counting a number of rejects, wherein one of said plural different discriminant functions is selected in accordance with the number counted by said reject counting step.
Claim 41
Independent Claim
45 words
A method for recognizing an unknown pattern comprising the steps of: selecting one of first and second dictionaries based on characteristics of the unknown pattern; and comparing the unknown pattern with a known pattern stored in the selected one of the first and second dictionaries.
Claim 42
Dependent Claim
44 words
A method according to Claim 41, wherein said selecting step selects the first dictionary in the case where the unknown pattern is a good quality pattern and wherein said selecting step selects said second dictionary in the case where said unknown pattern is degraded.
Claim 43
Dependent Claim
49 words
A method according to Claim 41, further comprising the step of counting a number of rejects, wherein said selecting step initially selects said first dictionary and, in the case where said reject counting step counts greater than a predetermined number of rejects, said selecting step selects said second dictionary.
Claim 44
Dependent Claim
45 words
A pattern recognition method comprising the steps of: inputting image information; segmenting the image information into pattern segments; selecting one of plural recognition modes in accordance with characteristics of the pattern segment; and providing a candidate for a pattern segment using the selected recognition mode.
Claim 45
Dependent Claim
33 words
A method according to Claim 44, further comprising the step of connecting a host to a network bus, and wherein said host is responsive to commands on said network bus for pattern recognition.
Claim 46
Dependent Claim
42 words
A method according to Claim 45, further comprising the steps of counting a number of rejects, initially selecting a first mode and, in a case where said reject counting step counts greater than a predetermined number of rejects, selecting said second mode.
Claim 47
Dependent Claim
39 words
A method according to Claim 44, further comprising the step of discriminating between character patterns and non-character patterns in said image information, and providing a candidate in the case where a character pattern is discriminated in said discriminating step.
Claim 48
Dependent Claim
26 words
A method according to Claim 44, further comprising the steps of processing character information, and previewing pattern recognition for character information processed in said processing step.
Claim 49
Dependent Claim
70 words
A pattern recognition method comprising the steps of: inputting image information on an image input device connected to a local area network bus; segmenting image information input by said image input device into segmented patterns; providing a segmented pattern to a pattern recognizer having plural recognition modes; selecting one of said plural recognition modes in accordance with characteristics of the unknown pattern; and providing a candidate for the unknown pattern.
Claim 50
Dependent Claim
49 words
A method according to Claim 49, further comprising the step of counting a number of rejects, wherein said selecting step initially selects a first mode and, in a case where said reject counting step counts greater than a predetermined number of rejects, said selecting step selects a second mode.
Claim 51
Dependent Claim
39 words
A method according to Claim 49, further comprising the steps of discriminating between character patterns and non-character patterns in said image information, and providing a candidate in the case where a character pattern is discriminated in said discriminating step.
Claim 52
Dependent Claim
27 words
A method according to Claim 49, further comprising the steps of processing character information, and previewing pattern recognition for character information processed in said information processing step.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.