Methods, systems, and media for evaluating images

Patent metadata

Field	Value
Publication number	US-11120252-B2
Application number	US-201916455824-A
Country	US
Kind code	B2
Filing date	Jun 28, 2019
Priority date	Dec 28, 2016
Publication date	Sep 14, 2021
Grant date	Sep 14, 2021

Abstract

A method may include obtaining an image including a face. The method may further include determining at least one time domain feature related to the face in the image and at least one frequency domain information related to the face in the image. The method may further include evaluating the quality of the image based on the at least one time domain feature and the frequency domain information.

First claim

What is claimed is:

1. A face recognition electronic system, comprising:
one or more image obtaining units configured to obtain an image;
one or more image processing units connected to the one or more image obtaining units, wherein during operation the one or more image processing units:
obtain an image including a face;
determine at least one time domain feature related to the face in the image, the at least one time domain feature related to the face indicating one or more features of the face presented in the image at a time spot and including a rotational angle of the face with respect to an axis or a confidence value of the face indicating whether the face is occluded;
determine at least one frequency domain information related to the face in the image; and
evaluate a quality of the image based on the at least one time domain feature and the frequency domain information.

2. The system of claim 1 , wherein to evaluate the quality of the image, the one or more image processing units:
determine whether the at least one time domain feature satisfies a condition;
in response to a determination that the at least one time domain feature satisfies the condition, evaluate the quality of the image based on the frequency domain information.

3. The system of claim 1 , wherein the at least one time domain feature related to the face includes at least one of a posture of the face, or a definition of the face in the image.

4. The system of claim 3 , wherein to determine the definition of the face in the image, the one or more image processing units:
extract at least one landmark point in the image, the at least one landmark point being with respect to a region of interest of the face;
determine at least one neighboring point of the at least one landmark point; and
determine, based on the at least one landmark point and the at least one neighboring point, a local variance related to the at least one landmark point, an average gradient related to the at least one landmark point, or an edge width related to the at least one landmark point in the image.

5. The system of claim 3 , wherein to evaluate the quality of the image, the one or more image processing units:
determine a value of a first parameter, the first parameter related to the definition of the face;
determine a value of a second parameter, the second parameter related to the posture of the face; and
determine a quality evaluation value of the image based on the value of the first parameter and the value of the second parameter.

6. The system of claim 1 , wherein to determine the at least one frequency domain information of the image, the one or more image processing units:
perform the time-to-frequency-domain transformation on one or more pixel values of the image;
weight the time-to-frequency-domain transformed one or more pixel values of the image based on a weighting matrix; and
determine the at least one frequency domain information of the image based on the weighted pixel values of the image.

7. A method implemented on at least one image processing electronic device, the method comprising:
obtaining an image including a face;
determining at least one time domain feature related to the face in the image, the at least one time domain feature related to the face indicating one or more features of the face presented in the image at a time spot and including a rotational angle of the face with respect to an axis or a confidence value of the face indicating whether the face is occluded;
determining at least one frequency domain information related to the face in the image; and
evaluating a quality of the image based on the at least one time domain feature and the frequency domain information.

8. The method of claim 7 , wherein the evaluating of the quality of the image includes:
determining whether the at least one time domain feature satisfies a condition;
in response to a determination that the at least one time domain feature satisfies the condition, evaluating the quality of the image based on the frequency domain information.

9. The method of claim 7 , wherein the at least one time domain feature related to the face includes at least one of a posture of the face, or a definition of the face in the image.

10. The method of claim 9 , wherein the determining of the definition of the face in the image includes:
extracting at least one landmark point in the image, the at least one landmark point being with respect to a region of interest of the face;
determining at least one neighboring point of the at least one landmark point; and
determining, based on the at least one landmark point and the at least one neighboring point, a local variance related to the at least one landmark point, an average gradient related to the at least one landmark point, or an edge width related to the at least one landmark point in the image.

11. The method of claim 9 , wherein the evaluating of the quality of the image includes:
determining a value of a first parameter, the first parameter related to the definition of the face;
determining a value of a second parameter, the second parameter related to the posture of the face; and
determining a quality evaluation value of the image based on the value of the first parameter and the value of the second parameter.

12. The method of claim 7 , wherein the determining of the at least one frequency domain information of the image includes:
performing the time-to-frequency-domain transformation on one or more pixel values of the image;
weighting the time-to-frequency-domain transformed one or more pixel values of the image based on a weighting matrix; and
determining the at least one frequency domain information of the image based on the weighted pixel values of the image.

Assignees

• Zhejiang Dahua Technology Co

Inventors

• Guo Siyu
• Wang Haiyang
• Hao Jingsong
• Wang Gang

Classifications

• G06V10/467

  Encoded features or binary features, e.g. local binary patterns [LBP] · CPC title

• G06V40/161 Primary

  Detection; Localisation; Normalisation · CPC title

• G06V40/167

  using comparisons between temporally consecutive images · CPC title

• G06V10/993

  Evaluation of the quality of the acquired pattern · CPC title

• G06V10/431

  Frequency domain transformation; Autocorrelation · CPC title