WO2014178105A1 - Dispositif d'estimation d'attributs - Google Patents

Dispositif d'estimation d'attributs Download PDF

Info

Publication number
WO2014178105A1
WO2014178105A1 PCT/JP2013/062575 JP2013062575W WO2014178105A1 WO 2014178105 A1 WO2014178105 A1 WO 2014178105A1 JP 2013062575 W JP2013062575 W JP 2013062575W WO 2014178105 A1 WO2014178105 A1 WO 2014178105A1
Authority
WO
WIPO (PCT)
Prior art keywords
attribute
user
image
database
attribute estimation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/JP2013/062575
Other languages
English (en)
Japanese (ja)
Inventor
康行 伊原
光洋 藤田
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
NEC Solution Innovators Ltd
Original Assignee
NEC Solution Innovators Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by NEC Solution Innovators Ltd filed Critical NEC Solution Innovators Ltd
Priority to PCT/JP2013/062575 priority Critical patent/WO2014178105A1/fr
Priority to JP2015514706A priority patent/JP5965057B2/ja
Publication of WO2014178105A1 publication Critical patent/WO2014178105A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/172—Classification, e.g. identification
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/178—Human faces, e.g. facial parts, sketches or expressions estimating age from face image; using age information for improving recognition

Definitions

  • the present invention relates to an attribute estimation device.
  • a device such as a cigarette vending machine or a bank ATM that requires personal identification
  • a device that captures a human face with a camera to acquire a facial image and estimates attributes such as age and gender from the facial image It has been proposed (for example, Patent Document 1).
  • the estimation of attributes such as age is performed by model learning in which a machine learns from a large number of model image information to create a model, and an attribute estimation target image is compared with the model.
  • an object of the present invention is to provide an attribute estimation device capable of estimating an attribute according to an evaluation criterion of the attribute estimation target person (user).
  • an attribute estimation apparatus of the present invention includes an image acquisition unit, an image information database, an attribute information database, a model learning unit, and an attribute estimation unit, and the image acquisition unit acquires an image of a user.
  • the image information database includes the acquired user image and a non-user image database
  • the attribute information database includes a user-assigned attribute database
  • the model learning means includes the non-user image. Extracting image features obtained from an information database, creating a self-evaluation standard model from the features with reference to the user-assigned attribute database
  • the attribute estimating means extracts features from the user image
  • a self-evaluation criterion attribute estimation result is generated from the feature with reference to the self-evaluation criterion model.
  • the attribute estimation device of the present invention it is possible to estimate an attribute based on an evaluation criterion of an attribute estimation target person (user).
  • FIG. 1 is a diagram showing an example of the configuration of the attribute estimation apparatus of the present invention.
  • FIG. 2 is a flowchart showing an example of a process flow in the attribute estimation apparatus of the present invention.
  • FIG. 3 is a flowchart showing an example of estimating attributes according to user evaluation criteria in the attribute estimation apparatus of the present invention.
  • FIG. 4 is a flowchart illustrating an example of estimating an attribute using an evaluation criterion other than the user in the attribute estimation device of the present invention.
  • FIG. 5 is a diagram showing an example of an annotation user interface in the attribute estimation apparatus of the present invention.
  • FIG. 6 is a diagram showing an example of a user interface for self-evaluation criteria analysis in the attribute estimation apparatus of the present invention.
  • FIG. 7 is a diagram showing an example of a user interface for analyzing other person evaluation criteria in the attribute estimation apparatus of the present invention.
  • the attribute information database further includes a non-user grant attribute database
  • the model learning unit further refers to the non-user grant attribute database and evaluates the other person's evaluation criteria from the features. It is preferable that the model is created, and the attribute estimation unit further generates an other person evaluation standard attribute estimation result from the feature with reference to the other person evaluation standard model. According to this aspect, in addition to the attribute estimation based on the self-evaluation criteria, the attribute estimation based on the evaluation criteria other than the user (others) can be performed.
  • the attribute estimation means includes a user interface creation means, and the attribute estimation result is displayed on the user interface.
  • the attribute estimation apparatus of the present invention further includes attribute assigning means, wherein the attribute assigning means indicates the user image other than the user and inputs the attribute, and the input attribute is other than the user A mode in which it is stored in the assigned attribute database is preferable.
  • attribute assigning means indicates the user image other than the user and inputs the attribute
  • the input attribute is other than the user A mode in which it is stored in the assigned attribute database is preferable.
  • the attribute assigning unit includes a user interface creating unit, and the image is displayed and the attribute is input in the user interface.
  • the attribute estimation apparatus of the present invention includes a user terminal and an attribute estimation information providing server, wherein the user terminal and the attribute information providing server are connectable via a communication network, and the user terminal
  • the attribute information providing server may include an acquisition unit, and the attribute information providing server may include an image information database, an attribute information database, a model learning unit, and an attribute estimation unit.
  • the communication network is not a constituent requirement of the attribute estimation device of the present invention.
  • the attribute estimation information providing server of the present invention can be connected to the user terminal of the attribute estimation apparatus of the present invention via a communication network, and includes an image acquisition means, an image information database, an attribute information database, a model learning means, and an attribute estimation means.
  • the image acquisition means acquires an image of a user;
  • the image information database includes an image database of the acquired user and a non-user image database, the attribute information database includes a user grant attribute database,
  • the model learning means extracts image features obtained from the non-user image information database, creates a self-evaluation reference model from the features with reference to the user-assigned attribute database, and the attribute estimation means includes: A feature is extracted from the user's image, and a self-evaluation criterion attribute estimation result is generated from the feature with reference to the self-evaluation criterion model.
  • the attribute information database further includes a non-user grant attribute database
  • the model learning unit further refers to the non-user grant attribute database and evaluates others based on the characteristics. It is preferable that a reference model is created, and the attribute estimation unit further generates an other person evaluation reference attribute estimation result from the feature with reference to the other person evaluation reference model.
  • the attribute estimation unit includes a user interface creation unit, and the attribute estimation result is displayed on the user interface.
  • the server further includes an attribute assigning unit, and the attribute assigning unit displays an image of the user other than the user and inputs the attribute, and the input attribute is the user It is preferable that the data is stored in a non-given attribute database.
  • the attribute assigning unit includes a user interface creating unit, and the image is displayed and the attribute is input in the user interface.
  • FIG. 1 shows a schematic diagram of an example of the attribute estimation apparatus of the present invention.
  • the attribute estimation apparatus of this example includes a user terminal 1 and an attribute estimation information providing server 3, both of which are connected by a communication network 2.
  • the user terminal 1 is not particularly limited, and examples thereof include a personal computer (PC), a mobile phone, and a smartphone.
  • the user terminal 1 preferably has a function of capturing an image of the user.
  • a PC with a camera, a mobile phone with a camera, and a smartphone with a camera are preferable.
  • a user interface (UI) described later is displayed on the user terminal 1.
  • the image is not particularly limited, and may be an image of a part of the body such as a face image or an image of the entire body.
  • attributes to be estimated include, for example, age, sex, facial expression, makeup, facial impression, clothing impression, and the like.
  • the communication line network 2 may be wired or wireless, and examples thereof include the Internet and a LAN.
  • the attribute estimation information providing server 3 is not particularly limited, and a general server can be used.
  • the attribute estimation information providing server 3 includes an image information database (DB) 31, an attribute information DB 32, a model learning unit 33, and an attribute estimation unit 34.
  • the image information DB 31 includes a non-user image DB 311 and a user image 312 is captured
  • the attribute information DB 32 includes a user grant attribute DB 321 and a non-user grant attribute DB 322.
  • the model learning means 33 includes an image feature extraction processing unit 331 and a model learning processing unit 332
  • the attribute estimation means 34 includes an image feature extraction processing unit 341 and an attribute estimation processing unit 342.
  • FIG 2 shows an example of the flow of attribute estimation processing by the attribute estimation apparatus of this example.
  • a user accesses (logs in) the attribute estimation information providing server 3 of the present example from the Internet using a camera-equipped smartphone (user terminal 1), thereby starting use of the device of the present example ( S1).
  • the user captures a face image (S2) and transmits the image to the apparatus of this example, the image information DB is updated in the apparatus of this example (S3).
  • an annotation UI is created by the attribute assigning means (S4) and displayed on the screen of a terminal such as a smartphone other than the user (others) (S5).
  • An example of the annotation UI is shown in FIG.
  • the annotation UI displays an image of another person, an annotation question ("Is it cute?" In the figure), and an answer icon ("No" in the figure) ”-2 to“ Yes ”+2 five-level evaluation) is displayed, and when someone other than the user answers, the other person will annotate the user's image, and the result is given to those other than the user. Accumulated (updated) in the attribute DB 321 (S6).
  • the user annotates the face image of another person other than the user, and the result is accumulated (updated) in the user assignment attribute DB 322 (S6).
  • Such processing (the portion surrounded by the dotted line in FIG. 2) is repeated for each image.
  • the image feature extraction processing unit 331 extracts image features from the non-user image DB, and the extracted image features and users Based on the attribute information from the assigned attribute DB 322, a self-evaluation reference model 333 is created (S7).
  • the attribute estimation means 34 the feature is extracted from the user image 312 in the image feature extraction processing unit 341, and the attribute estimation processing unit 342 refers to the model 333 of the self-evaluation criteria, and An attribute is estimated from the feature, and an estimation result 343 is generated.
  • a self-evaluation criterion analysis UI is created (S8), and attribute estimation results are displayed by the UI (S9).
  • An example of the UI of the attribute estimation result of the self-evaluation criteria is shown in FIG.
  • the face image of the user is displayed and the attribute estimation result of the self-evaluation criterion is displayed.
  • the face attribute is “impression” of two items “cute” and “healthy”
  • the attribute estimation result is “impression degree” of two items “cute” and “healthy”.
  • Each is displayed with a score out of 100.
  • the user determines whether or not to display the attribute estimation result based on the other person evaluation criteria (S10). When not displaying (No), it is complete
  • the image feature extraction processing unit 331 extracts image features from the non-user image DB 311, and the extracted image features and users Based on the attribute information from the non-giving attribute DB 321, the other person evaluation reference model 333 is created (updated) (S ⁇ b> 11). Then, as shown in FIG. 4, in the attribute estimation means 34, the feature is extracted from the user image 312 in the image feature extraction processing unit 341, and the feature estimation unit 342 refers to the other person-based model 333 and the feature is extracted. , The attribute is estimated, and an estimation result 343 is generated.
  • FIG. 7 An example of the UI of the attribute estimation result of the other person evaluation criterion is shown in FIG. 7 in the UI.
  • the face image of the user himself / herself is displayed and the attribute estimation result of the other person evaluation standard is displayed.
  • the attribute of the face is “impression” of two items “cute” and “healthy”
  • the attribute estimation result is “impression degree” of two items “cute” and “healthy”. Each is displayed with a score out of 100.
  • a process will be complete
  • the present invention is widely used in the field of estimating attributes such as age, gender, and facial impression degree from an image, and its application is not limited.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

La présente invention a pour objectif de proposer un dispositif d'estimation d'attributs adapté pour estimer des attributs d'après des critères d'évaluation définis par une personne (l'utilisateur) dont des attributs doivent être estimés, et collecter des images faciales et des informations d'attribut auprès de plusieurs utilisateurs. Un dispositif d'estimation d'attributs selon la présente invention est caractérisé en ce qu'il comprend des moyens d'acquisition d'image, une base de données d'informations d'image, une base de données d'informations d'attributs, des moyens d'apprentissage de modèle, et des moyens d'estimation d'attributs. Les moyens d'acquisition d'image obtiennent une image d'un utilisateur ; la base de données d'informations d'image contient l'image obtenue de l'utilisateur ainsi qu'une base de données d'images de personnes autres que l'utilisateur ; la base de données d'informations d'attributs contient une base de données d'attributs fournie par l'utilisateur ; les moyens d'apprentissage de modèle extraient des caractéristiques, d'une image obtenue dans la base de données d'images de personnes autres que l'utilisateur, et génèrent un modèle de critères d'auto-évaluation à partir des caractéristiques, en se référant à la base de données d'attributs fournie par l'utilisateur ; et les moyens d'estimation d'attributs extraient des caractéristiques, de l'image de l'utilisateur, et génèrent des résultats d'estimation d'attributs basés sur des critères d'auto-évaluation à partir des caractéristiques, en se référant au modèle de critères d'auto-évaluation.
PCT/JP2013/062575 2013-04-30 2013-04-30 Dispositif d'estimation d'attributs Ceased WO2014178105A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/JP2013/062575 WO2014178105A1 (fr) 2013-04-30 2013-04-30 Dispositif d'estimation d'attributs
JP2015514706A JP5965057B2 (ja) 2013-04-30 2013-04-30 属性推定装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2013/062575 WO2014178105A1 (fr) 2013-04-30 2013-04-30 Dispositif d'estimation d'attributs

Publications (1)

Publication Number Publication Date
WO2014178105A1 true WO2014178105A1 (fr) 2014-11-06

Family

ID=51843258

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2013/062575 Ceased WO2014178105A1 (fr) 2013-04-30 2013-04-30 Dispositif d'estimation d'attributs

Country Status (2)

Country Link
JP (1) JP5965057B2 (fr)
WO (1) WO2014178105A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108985133A (zh) * 2017-06-01 2018-12-11 北京中科奥森数据科技有限公司 一种人脸图像的年龄预测方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05266144A (ja) * 1991-03-20 1993-10-15 Hitachi Ltd デ−タ処理システム及び方法
JP2000285151A (ja) * 1999-03-31 2000-10-13 Toyota Motor Corp デザイン評価装置及びデザイン作成装置並びに方法
JP2002158870A (ja) * 2000-06-23 2002-05-31 Eastman Kodak Co 画像の顕著性及びアピール性に基づいて写真印画の数、寸法、及び、倍率を変更する方法
JP2007052575A (ja) * 2005-08-17 2007-03-01 Konica Minolta Holdings Inc メタデータ付与装置およびメタデータ付与方法
WO2012132418A1 (fr) * 2011-03-29 2012-10-04 パナソニック株式会社 Dispositif d'estimation de caractéristique

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05266144A (ja) * 1991-03-20 1993-10-15 Hitachi Ltd デ−タ処理システム及び方法
JP2000285151A (ja) * 1999-03-31 2000-10-13 Toyota Motor Corp デザイン評価装置及びデザイン作成装置並びに方法
JP2002158870A (ja) * 2000-06-23 2002-05-31 Eastman Kodak Co 画像の顕著性及びアピール性に基づいて写真印画の数、寸法、及び、倍率を変更する方法
JP2007052575A (ja) * 2005-08-17 2007-03-01 Konica Minolta Holdings Inc メタデータ付与装置およびメタデータ付与方法
WO2012132418A1 (fr) * 2011-03-29 2012-10-04 パナソニック株式会社 Dispositif d'estimation de caractéristique

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
NAOYUKI MIYAMOTO ET AL.: "Estimation of One's Subjective Age Using Facial Images", THE TRANSACTIONS OF THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS, vol. J90-A, no. 3, 1 March 2007 (2007-03-01), pages 240 - 247 *
SHIGERU AKAMATSU: "Recognition of Facial Expressions by Human and Computer [II", THE JOURNAL OF THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS, vol. 85, no. 10, 1 October 2002 (2002-10-01), pages 766 - 771 *
YUMI JINNOUCHI ET AL.: "Objective Age in Subjective Age Estimation System Using Facial Images", ITE TECHNICAL REPORT, vol. 30, no. 17, 25 February 2006 (2006-02-25), pages 13 - 14 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108985133A (zh) * 2017-06-01 2018-12-11 北京中科奥森数据科技有限公司 一种人脸图像的年龄预测方法及装置

Also Published As

Publication number Publication date
JPWO2014178105A1 (ja) 2017-02-23
JP5965057B2 (ja) 2016-08-03

Similar Documents

Publication Publication Date Title
US10028569B2 (en) Makeup application assistance device, makeup application assistance system, and makeup application assistance method
KR101978299B1 (ko) 콘텐츠 서비스 시스템에서의 콘텐츠 서비스를 위한 장치
EP3116199A1 (fr) Procédé de diffusion d'informations à base de dispositifs portés sur le corps et dispositif associé
WO2016145830A1 (fr) Procédé de traitement d'informations, terminal et support de stockage informatique
US20120313964A1 (en) Information processing apparatus, information processing method, and program
JPWO2015122195A1 (ja) 印象分析装置、ゲーム装置、健康管理装置、広告支援装置、印象分析システム、印象分析方法及びプログラム
JP7238902B2 (ja) 情報処理装置、情報処理方法、およびプログラム
CN110717399A (zh) 人脸识别方法和电子终端设备
JP6727776B2 (ja) 支援システム、支援方法、プログラム
US20150334189A1 (en) Method for matching multiple devices, and device and server system for enabling matching
JP2017033413A (ja) ロボット撮影システムおよび画像管理方法
JP6367169B2 (ja) 画像処理装置、画像処理方法、プログラムおよび記録媒体
JP5924114B2 (ja) 情報処理装置、情報処理方法、コンピュータプログラムおよび画像表示装置
JP2018195164A (ja) 解析装置、解析プログラム及び解析方法
JP5965057B2 (ja) 属性推定装置
CN110673737A (zh) 基于智能家居操作系统的显示内容的调整方法及装置
WO2017175447A1 (fr) Appareil de traitement d'informations, procédé de traitement d'informations, et programme
JP7457496B2 (ja) データ管理装置、データ管理方法およびデータ管理プログラム
JP6338191B2 (ja) 属性推定システム
JP2020052505A (ja) 健康状態判定システム、健康状態判定装置、サーバ、健康状態判定方法、及びプログラム
CN107862727B (zh) 健康信息推送方法和装置
US20240428902A1 (en) Information processing device and information processing method
JP7229626B1 (ja) プログラム、情報処理装置、情報処理システム、情報処理方法
JP2026070910A (ja) システム
JP2026070997A (ja) システム

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13883644

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2015514706

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13883644

Country of ref document: EP

Kind code of ref document: A1