IP Library Granted Patent US 12,411,889
Granted Patent B2
US 12,411,889 · App. 17/922,826 · Granted Sep 9, 2025

Image selection apparatus, image selection method, and non-transitory computer-readable medium

Inventors: Noboru Yoshida (Tokyo, JP); Yadong Pan (Tokyo, JP); Ryo Kawai (Tokyo, JP); Jianquan Liu (Tokyo, JP); Shoji Nishimura (Tokyo, JP)
Assignee: NEC CORPORATION
G06F16/7328G06V20/41G06V20/48G06V20/52G06V10/945
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,411,889
App. No.
17/922,826
Granted
Sep 9, 2025
Kind
B2
Abstract

A query acquisition unit ( 610 ) acquires a query video. A query frame selection unit ( 620 ) selects a plurality of query frames from among a plurality of first frame images included in the query video. A video selection unit ( 630 ) selects a video by using the plurality of query frames. At this time, the video selection unit ( 630 ) uses references (1) and (2) below as a condition for selecting the video. Reference (1) A condition that a “similar frame image whose degree of similarity to a query frame satisfies a first reference is present” is satisfied for at least two query frames. Reference (2) When query frames associated with each of at least two similar frame images are arranged in the same order as that of the at least two similar frame images, the arrangement order coincides with an arrangement order of at least two query frames in a query video.

Claims (40)

1. An image selection apparatus comprising:

at least one memory storing instructions; and

at least one processor configured to execute the instructions to perform operations comprising:

acquiring a query video including a plurality of first frame images;

selecting a plurality of query frames from among the plurality of first frame images, wherein the plurality of query frames are selected by repeatedly performing a process of selecting, as a query frame following a first of the query frames, the first frame image having a change amount from the first query frame equal to or greater than a predetermined reference; and

selecting a video by using the plurality of query frames, based on at least first and second conditions for selecting the video, wherein

the first condition is that a similar frame image having a degree of similarity to the query frame satisfies a first reference is present for at least two of the query frames,

the second condition is that, in a case in which the query frames associated with each of at least two similar frame images are arranged in a same order as that of the at least two similar frame images, an arrangement order of the at least two similar frame images coincides with an arrangement order of the at least two query frames in the query video.

2. The image selection apparatus according to claim 1 , wherein

the query video includes a person, and

the operations comprise using, as the degree of similarity, a degree of similarity of a pose of the person.

3. The image selection apparatus according to claim 2 , wherein

a third condition that is to be satisfied by the query frame is that an information amount related to the person included in the query frame satisfies a second reference.

4. The image selection apparatus according to claim 3 , wherein

each of the plurality of first frame images includes, as information indicating a pose of the person, joint information indicating a position of a joint of the person, and

the second reference is that a number of joints included in the joint information satisfies a reference.

5. The image selection apparatus according to claim 1 , wherein

the operations comprise selecting at least one of the query frames according to an input from a user.

6. The image selection apparatus according to claim 1 , wherein

the operations comprise setting the predetermined reference by using an input from a user.

7. The image selection apparatus according to claim 1 , wherein

the operations comprise setting the third predetermined reference by using the plurality of first frame images.

8. The image selection apparatus according to claim 1 , wherein

the operations comprise selecting the plurality of query frames from among the plurality of first frame images at predetermined intervals.

9. The image selection apparatus according to claim 1 , wherein

the operations comprise selecting an n-th of the first frame images as a first of the query frames, and

n is an integer set in advance or by a user input.

10. An image selection method performed by a computer and comprising:

by a computer:

acquiring a query video including a plurality of first frame images;

selecting a plurality of query frames from among the plurality of first frame images, wherein the plurality of query frames are selected by repeating processing of selecting, as a query frame following a first of the query frames, the first frame image having a change amount from the first query frame equal to or more than a predetermined reference; and

selecting a video by using the plurality of query frames, based on at least first and second conditions selecting the video, wherein

the first condition is that a similar frame image having a degree of similarity to the query frame satisfies a first reference is present is satisfied for at least two of the query frames,

the second condition is that, in a case in which the query frames associated with each of at least two similar frame images are arranged in a same order as that of the at least two similar frame images, an arrangement order of the at least two similar frame images coincides with an arrangement order of the at least two query frames in the query video.

11. A non-transitory computer-readable medium storing a program executable by a computer to perform operations comprising:

acquiring a query video including a plurality of first frame images;

selecting a plurality of query frames from among the plurality of first frame images, wherein the plurality of query frames are selected by repeating processing of selecting, as a query frame following a first of the query frames, the first frame image having a change amount from the first query frame equal to or more than a predetermined reference; and

selecting a video by using the plurality of query frames, based on at least first and second conditions selecting the video, wherein

the first condition is that a similar frame image having a degree of similarity to the query frame satisfies a first reference is present is satisfied for at least two of the query frames,

the second condition is that, in a case in which the query frames associated with each of at least two similar frame images are arranged in a same order as that of the at least two similar frame images, an arrangement order of the at least two similar frame images coincides with an arrangement order of the at least two query frames in the query video.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Nov 2, 2022
From: YOSHIDA, NOBORU; PAN, YADONG; KAWAI, RYO; LIU, JIANQUAN; NISHIMURA, SHOJI
To: NEC CORPORATION
Reel/Frame 061628/0510 →
Continuity (1)
Related Publication 20230161815A1 · May 25, 2023
References Cited (17)
US 6859802B1 · Rui · 2005 [cited by examiner]
US 20020174120A1 · Zhang · 2002 [cited by examiner]
US 20130028517A1 · Yoo et al. · 2013 [cited by applicant]
US 20160110453A1 · Kim · 2016 [cited by examiner]
US 20180276246A1 · Matsumoto · 2018 [cited by applicant]
US 20190147292A1 · Watanabe et al. · 2019 [cited by applicant]
US 20210058667A1 · He · 2021 [cited by examiner]
CN 110728209A · 2020 [cited by applicant]
JP 2001134589A · 2001 [cited by applicant]
JP 2006260405A · 2006 [cited by applicant]
JP 2011118790A · 2011 [cited by applicant]
JP 2014522035A · 2014 [cited by applicant]
JP 2017117302A · 2017 [cited by applicant]
JP 2019091138A · 2019 [cited by applicant]
WO 2006025272A1 · 2006 [cited by applicant]
International Search Report for PCT Application No. PCT/JP2020/019255, mailed on Jul. 14, 2020. [cited by applicant]
Zhe Cao, Tomas Simon, Shih-En Wei, Yaser Sheikh, “Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields”, The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, p. 7291-7299. [cited by applicant]