Patent US9275270B2 successfully refreshed from Google Patents.
← Back to Patent Explorer
US9275270B2 Active / In Force US ⏱️ 6.6 years remaining

Information processing apparatus and control method thereof

Assignee: Canon Inc
•
Filing: 2013-06-20
•
Granted: 2016-03-01
•
Expires: 2033-06-20
📄 Original PDF 🌐 Google Patents 🔄 Sync Now

Abstract

Abstract This invention provides a technique which can enhance personal recognition precision in personal recognition processing of a face in an image. To this end, a management unit classifies feature patterns each including feature information of a plurality of parts of a face region of an object extracted from image data, and manages the feature patterns using a dictionary. A segmenting unit determines whether or not feature information of each part of the face region of the object is segmented, and segments the feature information of the part of interest into a plurality of feature information as new feature information. A registration unit registers a feature pattern as a combination of the new feature information of the part of interest and feature information of parts other than the part of interest in the dictionary as a new feature pattern of the object.

Inventors

  • 👤 Tetsuya Suwa
  • 👤 Hiroyasu Kunieda
  • 👤 Yusuke Hashii
  • 👤 Naoki Sumi
  • 👤 Takashi Nakamura
  • 👤 Takashi Fujita
  • 👤 Kiyoshi Umeda
  • 👤 Hiroyuki Sakai

Bibliographic Details

Application Number —
Jurisdiction US
Data Source google_patents
Last Synchronized 2026-11-09 09:47
CPC Classifications
G; G06; G06V; G06V40/00; G06V40/10; G06V40/16; G06V40/168; G06V40/171; G06K9/00268; G06K9/00281

Claims Analysis

Total Claims: 27
Claim 1
Independent Claim 170 words
An apparatus comprising: a management unit that classifies feature patterns for identifying an object, where each pattern includes a plurality of feature amounts corresponding to a plurality of parts of a face region of the object extracted from image data, and that manages the feature patterns using a dictionary; a dividing unit that divides a feature amount of a target part among the plurality of parts of an object of interest, into a plurality of division feature amounts as a plurality of new feature amounts of the target part; and a registration unit that registers, in a case where the dividing unit divides the feature amount of the target part into the plurality of new feature amounts, a plurality of new feature patterns for identifying the object of interest, where the plurality of new feature patterns are obtained by combining each of the plurality of new feature amounts of the target part divided by the dividing unit with feature amounts of parts other than the target part, in the dictionary.
Claim 2
Independent Claim 170 words
An apparatus comprising: a management unit that classifies feature patterns for identifying an object, where each pattern includes a plurality of feature amounts corresponding to a plurality of parts of a face region of the object extracted from image data, and that manages the feature patterns using a dictionary; a dividing unit that divides a feature amount of a target part among the plurality of parts of an object of interest, into a plurality of division feature amounts as a plurality of new feature amounts of the target part; and a registration unit that registers, in a case where the dividing unit divides the feature amount of the target part into the plurality of new feature amounts, a plurality of new feature patterns for identifying the object of interest, where the plurality of new feature patterns are obtained by combining each of the plurality of new feature amounts of the target part divided by the dividing unit with feature amounts of parts other than the target part, in the dictionary.
Claim 3
Dependent Claim 20 words
The apparatus according to claim 1 , wherein the dictionary includes not less than two feature patterns for respective objects.
Claim 4
Dependent Claim 82 words
The apparatus according to claim 2 , further comprising: a first determination unit that determine whether or not a frequency of occurrence of a feature patterns managed in the dictionary, which is managed by said management unit, in a plurality of image data upon recognition of a face region in image data using the dictionary satisfies a predetermined criterion; and an excluding unit that excludes a feature pattern which is determined by said first determination unit not to satisfy the predetermined criterion.
Claim 5
Dependent Claim 28 words
The apparatus according to claim 1 , wherein one of the feature patterns managed in the dictionary includes a feature amount of an accessory in a face region.
Claim 6
Dependent Claim 131 words
The apparatus according to claim 1 , further comprising: a second determination unit that determines, in a case where variations of feature patterns of an object managed in the dictionary exceed a pre-set value, whether or not variations of the feature patterns of the object managed in the dictionary and a new feature pattern of the object in the dictionary are smaller than the variations of the feature pattern of the object in the dictionary, wherein in a case where said second determination unit determines that the variations of the feature patterns of the object managed in the dictionary and the new feature pattern are smaller than the variations of the feature patterns of the object managed in the dictionary, said registration unit registers the new feature pattern in the dictionary.
Claim 7
Dependent Claim 97 words
The apparatus according to claim 1 , further comprising a specifying unit which specifies one or more peaks of the feature amount of the target part of the face region in a frequency distribution, wherein, in a case where the specifying unit specifies a plurality of peaks of the feature amount of the target part of the face region, the division unit divides the feature amount of the target part into the plurality of division feature amounts as the plurality of new feature amounts of the target part based on the peaks specified by the specifying unit.
Claim 8
Dependent Claim 69 words
The apparatus according to claim 6 , wherein, in a case where the specifying unit does not specify a plurality of peaks of the feature amount of the target part of the face region in a frequency distribution, the division unit does not divide the feature amount of the target part into the plurality of division feature amounts as the plurality of new feature amounts of the target part.
Claim 9
Dependent Claim 39 words
The apparatus according to claim 1 , wherein the registration unit re-registers the plurality of new feature patterns by replacing the plurality of the feature pattern having been managed in the dictionary with the plurality of new feature patterns.
Claim 10
Dependent Claim 23 words
The apparatus according to claim 1 , further comprising a face recognition unit which recognizes a face in an image using the dictionary.
Claim 11
Dependent Claim 25 words
The apparatus according to claim 1 , wherein the plurality of parts include at least two of an eye, an eyebrow, a mouth, and nose.
Claim 12
Dependent Claim 101 words
The apparatus according to claim 1 , further comprising a determining unit determining whether the number of peaks of the feature amount of the target part of the face region in a frequency distribution is more than one, wherein, in a case where the determining unit determines that the number of peaks of the feature amount of the target part of the face region in the frequency distribution is more than one, the dividing unit divides the feature amount of the target part into the plurality of division feature amounts as the plurality of new feature amounts of the target part.
Claim 13
Dependent Claim 48 words
The apparatus according to claim 1 , wherein the plurality of new feature patterns of the object are obtained by combining a first feature amount, a second feature amount and a third feature amount, each corresponding to a part different from each other of the object of interest.
Claim 14
Independent Claim 186 words
A method of controlling an information processing apparatus managing a plurality of images in which a face is included, comprising the steps of: classifying, with a classifying unit, feature patterns for identifying an object, where each pattern includes a plurality of feature amounts corresponding to a plurality of parts of a face region of the object extracted from image data, and managing the feature patterns using a dictionary; dividing, with a dividing unit a feature amount of a target part among the plurality of parts of an object interest into a plurality of division feature amounts as a plurality of new feature amounts of the target part; and registering, with a registering unit, in a case where the feature amount of the target part is divided into the plurality of new feature amounts, a plurality of new feature patterns for identifying the object of interest, where the plurality of new feature patterns are obtained by combining each of the plurality of new feature amounts of the target part divided in the dividing step with feature amounts of parts other than the target part in the dictionary.
Claim 15
Independent Claim 186 words
A method of controlling an information processing apparatus managing a plurality of images in which a face is included, comprising the steps of: classifying, with a classifying unit, feature patterns for identifying an object, where each pattern includes a plurality of feature amounts corresponding to a plurality of parts of a face region of the object extracted from image data, and managing the feature patterns using a dictionary; dividing, with a dividing unit a feature amount of a target part among the plurality of parts of an object interest into a plurality of division feature amounts as a plurality of new feature amounts of the target part; and registering, with a registering unit, in a case where the feature amount of the target part is divided into the plurality of new feature amounts, a plurality of new feature patterns for identifying the object of interest, where the plurality of new feature patterns are obtained by combining each of the plurality of new feature amounts of the target part divided in the dividing step with feature amounts of parts other than the target part in the dictionary.
Claim 16
Dependent Claim 20 words
The method according to claim 13 , wherein the dictionary includes not less than two feature patterns for respective objects.
Claim 17
Dependent Claim 65 words
The method according to claim 14 , further comprising the step of: determining whether or not a frequency of occurrence of a managed feature pattern in the dictionary in a plurality of image data upon recognition of a face region in image data using the dictionary satisfies a predetermined criterion, wherein a feature pattern which is determined not to satisfy the predetermined criterion, is excluded.
Claim 18
Dependent Claim 28 words
The method according to claim 13 , wherein one of the feature patterns managed in the dictionary includes a feature amount of an accessory in a face region.
Claim 19
Dependent Claim 123 words
The method according to claim 13 , further comprising: determining, in a case where variations of feature patterns of an object managed in the dictionary exceed a pre-set value, whether or not variations of the feature patterns of the object managed in the dictionary and a new feature pattern of the object in the dictionary are smaller than the variations of the feature patterns of the object managed in the dictionary, wherein, in a case where it is determined that the variations of the feature patterns of the object managed in the dictionary and the new feature pattern are smaller than the variations of the feature patterns of the object managed in the dictionary, the new feature pattern is registered in the dictionary.
Claim 20
Dependent Claim 168 words
A non-transitory computer-readable storage medium storing a program for controlling a computer to execute respective steps of a method, the method comprising the steps of: classifying feature patterns for identifying an object, where each pattern includes feature amounts corresponding to a plurality of parts of a face region of the object extracted from image data, and managing the feature patterns using a dictionary; dividing, a feature amount of a target part among the plurality of parts of an object of interest into a plurality of division feature amounts as a plurality of new feature amounts of the target part; and registering, in a case where the feature amount of the target part is divided into the plurality of new feature amounts, a plurality of new feature patterns for identifying the object of interest, where the plurality of feature patterns are obtained by combining each of the plurality of new feature amounts of the target part with feature amounts of parts other than the target part in the dictionary.
Claim 21
Dependent Claim 168 words
A non-transitory computer-readable storage medium storing a program for controlling a computer to execute respective steps of a method, the method comprising the steps of: classifying feature patterns for identifying an object, where each pattern includes feature amounts corresponding to a plurality of parts of a face region of the object extracted from image data, and managing the feature patterns using a dictionary; dividing, a feature amount of a target part among the plurality of parts of an object of interest into a plurality of division feature amounts as a plurality of new feature amounts of the target part; and registering, in a case where the feature amount of the target part is divided into the plurality of new feature amounts, a plurality of new feature patterns for identifying the object of interest, where the plurality of feature patterns are obtained by combining each of the plurality of new feature amounts of the target part with feature amounts of parts other than the target part in the dictionary.
Claim 22
Dependent Claim 94 words
The non-transitory computer-readable storage medium according to claim 18 , further comprising a step of specifying one or more peaks of a feature amount of target part of the face region in a frequency distribution, wherein, in a case where in the specifying step a plurality of peaks of the feature amount of the target part of the face are specified, the dividing step divides the feature amount of the target part into plurality of division feature amounts as the plurality of new feature amounts of the target part based on the specified peaks.
Claim 23
Dependent Claim 68 words
The non-transitory computer-readable storage medium according to claim 19 , wherein, in a case where in the specifying step a plurality of peaks of the feature amount of the target part of the face are not specified, the dividing step does not divide the feature amount of the target part into the plurality of division feature amounts as the plurality of new feature amounts of the target part.
Claim 24
Dependent Claim 66 words
The non-transitory computer-readable storage medium according to claim 18 , wherein, in a case where in the dividing step the feature amount of the one part is divided into the plurality of new feature amounts, the registering step re-registers the plurality of new feature patterns by replacing the plurality of the feature patterns having been managed in the dictionary with the plurality of new feature patterns.
Claim 25
Dependent Claim 26 words
The non-transitory computer-readable storage medium according to claim 18 , further comprising a face recognition step of recognizing a face in an image using the dictionary.
Claim 26
Dependent Claim 25 words
The non-transitory computer-readable storage medium according to claim 18 , wherein the plurality of parts include at least two of eye, eyebrow, mouth, and nose.
Claim 27
Dependent Claim 46 words
The non-transitory computer-readable storage medium according to claim 18 , wherein the plurality of new feature patterns of the object are obtained by combining a first feature amount, a second feature amount and a third feature amount, each corresponding to a part different from each other.

USPTO Annuity & Maintenance Windows

Annuity Stage Window Opens Payment Deadline Surcharge Period Until Estimated Fee Status
3.5 Years (Year 4) 2019-03-01 2019-08-31 2020-03-01 $2000.00 Window Closed / Lapsed
7.5 Years (Year 8) 2023-03-01 2023-08-31 2024-03-01 $3760.00 Window Closed / Lapsed
11.5 Years (Year 12) 2027-03-01 2027-08-31 2028-03-01 $7700.00 Pending Window

Assigned Categories & Tags

No categories assigned to this patent yet.

Assign Category Tag