Method for selecting image of interest to construct retrieval database and image control system performing the same

Patent metadata

Field	Value
Publication number	US-11921774-B2
Application number	US-202016796943-A
Country	US
Kind code	B2
Filing date	Feb 21, 2020
Priority date	Jul 19, 2019
Publication date	Mar 5, 2024
Grant date	Mar 5, 2024

Abstract

Exemplary embodiments relate to a method for selecting an image of interest to construct a retrieval database including receiving an image captured by an imaging device, detecting an object of interest in the received image, selecting an image of interest based on at least one of complexity of the image in which the object of interest is detected and image quality of the object of interest, and storing information related to the image of interest in the retrieval database, and an image control system performing the same.

First claim

What is claimed is:

1. A method for selecting an image of interest to construct a retrieval database, performed by a computing device including a processor, the method comprising:
receiving an image captured by an imaging device;
detecting an object of interest in the received image;
selecting an image of interest based on at least one of complexity of the image in which the object of interest is detected and image quality of the object of interest, wherein selecting the image of interest comprises calculating an image quality score of the object of interest based at least in part on a detection accuracy of the object of interest, wherein the detection accuracy of the object of interest is calculated based on the percent of the object of interest included in a region of interest, a ratio between background and foreground in the region of interest, and a blur of the region of interest; and
storing information related to the image of interest in the retrieval database,
wherein the selecting the image of interest further comprises:
calculating a complexity score of the image by analyzing the entire image in which the object of interest is detected,

wherein the complexity score increases as a total number of objects in the entire image increases,
wherein the image includes a first image and a second image received after the first image,
both the first image and the second image include the object of interest, and
the processor selects the second image as the image of interest when the first image has the complexity score that is equal to or higher than a first threshold and the second image has the complexity score that is lower than the first threshold and the image quality score that is equal to or higher than a second threshold.

2. The method for selecting an image of interest to construct a retrieval database according to claim 1 , wherein the selecting the image of interest further comprises:
comparing the complexity score of the image with the first threshold, and selecting an image having the complexity score that is equal to or higher than the first threshold as the image of interest.

3. The method for selecting an image of interest to construct a retrieval database according to claim 1 , wherein the selecting the image of interest comprises:
calculating the image quality score of the region of interest including at least part of the object of interest in the image in which the object of interest is detected; and
comparing the image quality score of the region of interest of the image with the second threshold, and selecting an image having the image quality score that is equal to or higher than the second threshold as the image of interest.

4. The method for selecting an image of interest to construct a retrieval database according to claim 1 , wherein the selecting the image of interest comprises:
comparing the complexity score of the image with the first threshold;
calculating the image quality score of the region of interest including at least part of the object of interest in the image in which the object of interest is detected;
comparing the image quality score of the region of interest of the image with the second threshold; and
selecting, as the image of interest, an image having the complexity score that is equal to or higher than the first threshold or an image having the image quality score that is equal to or higher than the second threshold.

5. The method for selecting an image of interest to construct a retrieval database according to claim 1 , wherein the complexity score of the image is calculated through at least one of a bottom-up approach using at least one of spatial information and temporal information of the image and a top-down approach using information related to the detected object of interest.

6. The method for selecting an image of interest to construct a retrieval database according to claim 1 , wherein the receiving the image comprises receiving a third image captured by a first imaging device and a fourth image captured by a second imaging device.

7. The method for selecting an image of interest to construct a retrieval database according to claim 6 , wherein the selecting the image of interest comprises selecting the image of interest based on complexity of the third image and image quality of the object of interest of the fourth image, when the object of interest is detected in the third image and the fourth image.

8. The method for selecting an image of interest to construct a retrieval database according to claim 6 , wherein the selecting the image of interest comprises selecting the image of interest based on complexity of the third image, image quality of the object of interest of the third image, complexity of the fourth image and image quality of the object of interest of the fourth image, when the object of interest is detected in the third image and the fourth image.

9. A non-transitory computer-readable recording medium that is readable by a computing device and stores program instructions executable by the computing device, and when the program instructions are executed by a processor of the computing device, the program instructions cause the processor to:
receive an image captured by an imaging device;
detect an object of interest in the received image;
select an image of interest based on at least one of complexity of the image in which the object of interest is detected and image quality of the object of interest, wherein selecting the image of interest comprises calculating an image quality score of the object of interest based at least in part on a detection accuracy of the object of interest, wherein the detection accuracy of the object of interest is calculated based on the percent of the object of interest included in a region of interest, a ratio between background and foreground in the region of interest, and a blur of the region of interest; and
store information related to the image of interest in a retrieval database,
wherein the selecting the image of interest further comprises:
calculating a complexity score of the image by analyzing the entire image in which the object of interest is detected,

wherein the complexity score increases as a total number of objects in the entire image increases,
wherein the image includes a first image and a second image received after the first image,
both the first image and the second image include the object of interest, and
the program instructions cause the processor to select the second image as the image of interest when the first image has the complexity score that is equal to or higher than a first threshold and the second image has the complexity score that is lower than the first threshold and the image quality score that is equal to or higher than a second threshold.

10. An image control system, comprising:
at least one imaging device that images an object and a background;
a server that selects, as an image of interest, an image having a possibility that a user will retrieve later the image received from the imaging device, wherein the image of interest is selected based at least in part on a complexity of the image, wherein selecting the image of interest comprises calculating an image quality score of an object of interest based at least in part on a detection accuracy of the object of interest, wherein the detection accuracy of the object of interest is calculated based on the percent of the object of interest included in a region of interest, a ratio between background and foreground in the region of interest, and a blur of the region of interest; and
a first database constructed using the image of interest,
wherein the selecting the image of interest further comprises:
cal

Assignees

• Korea Inst Sci & Tech

Inventors

• Kim Ig Jae
• Choi Heeseung
• Kim Haksub
• Chae Seungho
• Yang Yoonsik

Classifications

• G06F16/58 Primary

  Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually · CPC title

• G06F16/24578

  using ranking · CPC title

• G06F16/51

  Indexing; Data structures therefor; Storage structures · CPC title

• G06F16/532

  Query formulation, e.g. graphical querying · CPC title

• G06F16/5854

  using shape and object relationship · CPC title