Image extraction device, image extraction method, program, and recording medium

Patent metadata

Field	Value
Publication number	US-9870507-B2
Application number	US-201615135654-A
Country	US
Kind code	B2
Filing date	Apr 22, 2016
Priority date	Jun 22, 2015
Publication date	Jan 16, 2018
Grant date	Jan 16, 2018

Abstract

In the image extraction device, an instruction acquisition unit acquires an instruction input by a user, and an image group selection unit selects a second image group, which has a smaller number of images than a first image group, from the first image group in response to the instruction. Then, an extraction reference determination unit determines an image extraction reference when extracting an image from the second image group based on images included in the first image group, and an image extraction unit extracts one or more images, the number of which is smaller than the number of images in the second image group, from the second image group according to the image extraction reference.

First claim

What is claimed is:

1. An image extraction device, comprising:
an instruction acquisition unit that acquires an instruction input by a user;
an image group selection unit that selects a second image group, which has a smaller number of images than a first image group, from the first image group in response to the instruction;
an extraction reference determination unit that determines an image extraction reference when extracting an image from the second image group based on images included in the first image group;
an image extraction unit that extracts one or more images, the number of which is smaller than the number of images in the second image group, from the second image group according to the image extraction reference; and
an image analysis unit that analyzes images included in the first image group and calculates a relationship between persons present in the images included in the first image group based on the analysis result;
the image analysis unit comprising:
an image information acquisition section that acquires image information regarding images included in the first image group;
an image feature analysis section that analyzes features of the images included in the first image group based on the image information; and
a relationship calculation section that calculates a relationship between persons present in the images included in the first image group based on the image information and the image feature, wherein

the extraction reference determination unit determines the image extraction reference according to the relationship between persons,
the relationship calculation section determines a degree of importance of each person from the relationship between persons,
the extraction reference determination unit determines the image extraction reference according to the degree of importance of each person,
the relationship calculation section determines one or more persons whose number of times of appearance is equal to or greater than a threshold value in the first image group, among persons present in the images included in the first image group, as main persons, and determines one or more persons who are present in the images included in the first image group and whose distances from the main persons are equal to or less than a threshold value, among persons other than the main persons who are present in the images included in the first image group, as important persons, and
the extraction reference determination unit determines the image extraction reference according to the main persons and the important persons.

2. The image extraction device according to claim 1 ,
wherein the image analysis unit further calculates a relationship between persons present in images included in the second image group based on an analysis result of the images included in the second image group, and
the extraction reference determination unit further determines the image extraction reference according to the relationship between persons present in the images included in the second image group.

3. The image extraction device according to claim 2 ,
wherein the extraction reference determination unit determines the image extraction reference by giving priority to the relationship between persons present in the images included in the second image group over the relationship between persons present in the images included in the first image group.

4. The image extraction device according to claim 1 ,
wherein the relationship calculation section determines one or more persons whose number of times present in a center position of each image included in the first image group is equal to or greater than a threshold value, among persons whose number of times of appearance in the first image group is equal to or greater than a threshold value, as the main persons.

5. The image extraction device according to claim 1 ,
wherein the relationship calculation section determines one or more persons whose number of times present toward a front direction in each image included in the first image group is equal to or greater than a threshold value, among persons whose number of times of appearance in the first image group is equal to or greater than a threshold value, as the main persons.

6. The image extraction device according to claim 1 ,
wherein the relationship calculation section determines one or more persons whose number of times present with a face size equal to or greater than a threshold value in each image included in the first image group is equal to or greater than a threshold value, among persons whose number of times of appearance in the first image group is equal to or greater than a threshold value, as the main persons.

7. The image extraction device according claim 1 ,
wherein the relationship calculation section further determines two or more persons other than the main persons, who are present in a center position of each image included in the first image group, as important persons.

8. The image extraction device according to claim 1 ,
wherein the relationship calculation section further determines two or more persons other than the main persons, who are present toward a front direction in each image included in the first image group, as important persons.

9. The image extraction device according to claim 1 ,
wherein the relationship calculation section further determines two or more persons other than the main persons, who are present with a face size equal to or greater than a threshold value in each image included in the first image group, as important persons.

10. The image extraction device according to claim 1 ,
wherein the relationship calculation section further calculates a relationship between a person present in each image included in the first image group and an object other than the person, and
the extraction reference determination unit further determines the image extraction reference according to the relationship between the person and the object present in each image included in the first image group.

11. The image extraction device according to claim 10 ,
wherein the relationship calculation section determines a degree of importance of the object from the relationship between the person and the object, and
the extraction reference determination unit determines the image extraction reference according to the degree of importance of the object.

12. The image extraction device according to claim 11 ,
wherein the relationship calculation section determines a scene that has been imaged a number of times equal to or greater than a threshold value, among scenes of images included in the first image group, as an important scene, and determines a person and an object, which are present in the images included in the first image group and whose relationship with the important scene is equal to or greater than a threshold value, as an important person and an important object, and
the extraction reference determination unit determines the image extraction reference according to the main person, the important person, and the important object.

13. The image extraction device according to claim 11 ,
wherein the relationship calculation section determines an imaging date on which imaging has been performed a number of times equal to or greater than a threshold value, among imaging dates of images included in the first image group, as an important imaging date, and determines a person, an object, and a scene which are present in the images included in the first image group and whose relationship with the important imaging date is equal to or greater than a threshold value, as an important person, an important object, an

Assignees

• Fujifilm Corp

Inventors

• Yamaji Kei
• Asada Junichi

Classifications

• G06F17/30247

  Physics · mapped topic

• G06K9/00677 Primary

  Physics · mapped topic

• G06V20/30 Primary

  in albums, collections or shared content, e.g. social network photos or video · CPC title

• G06F16/583 Primary

  using metadata automatically derived from the content · CPC title