Method and device for verifying image and video

Patent metadata

Field	Value
Publication number	US-11625819-B2
Application number	US-202017073880-A
Country	US
Kind code	B2
Filing date	Oct 19, 2020
Priority date	Oct 18, 2019
Publication date	Apr 11, 2023
Grant date	Apr 11, 2023

Abstract

A method for verifying an image can include: acquiring a first feature point set of a source image and a second feature point set of a target image; determining a target local feature point pair based on the first feature point set and the second feature point set; determining a mapped point of the first feature point on the target image; determining a distance between a second feature point and the mapped point; acquiring a quantity of reference local feature point pairs; and determining that the target image is an image acquired by copying the source image based on the quantity being greater than a target quantity.

First claim

What is claimed is:

1. A method for verifying an image, comprising:
acquiring a first feature point set of a source image and a second feature point set of a target image;
determining a target local feature point pair based on the first feature point set and the second feature point set, wherein the target local feature point pair comprises a first feature point in the source image and a second feature point in the target image;
determining a mapped point of the first feature point on the target image;
determining a distance between the second feature point and the mapped point;
acquiring a quantity of reference local feature point pairs, wherein the reference local feature point pairs are target local feature point pairs with distances being less than a target distance threshold; and
determining that the target image is an image acquired by copying the source image based on the quantity being greater than a target quantity,
wherein said determining the mapped point of the first feature point on the target image comprises:
determining a rigid body transformation matrix based on coordinates of the first feature point in the source image and coordinates of the second feature point in the target image;
acquiring a homography matrix based on the rigid body transformation matrix; and
determining the mapped point based on the homography matrix, wherein the mapped point is formed by mapping the first feature point to the target image.

2. The method according to claim 1 , wherein said determining the target feature point pair based on the first feature point set and the second feature point set comprises:
determining distances between any local feature point in the first feature point set and a plurality of reference feature points, wherein the reference feature points are local feature points in the second feature point set;
acquiring a first distance and a second distance, wherein the first distance is a distance between the local feature point and a closest reference feature point, and the second distance is a distance between the local feature point and a second closest reference feature point; and
determining, based on a ratio of the first distance to the second distance being less than a target ratio, the reference feature point corresponding to the first distance and the local feature point as the target local feature point pair.

3. The method according to claim 1 , wherein said acquiring the quantity of the reference local feature point pairs comprises:
searching for local feature points in the source image or the target image based on a target sliding window;
determining the reference local feature point pairs based on the target local feature point pairs in the target sliding window; and
counting the quantity of the reference local feature point pairs.

4. A method for verifying a video, comprising:
acquiring a plurality of matched image pairs between a source video and a target video, wherein each of matched image pairs comprises a source image in the source video and a target image in the target video;
acquiring a first feature point set of the source image and a second feature point set of the target image in the each of the matched image pairs;
determining a target local feature point pair based on the first feature point set and the second feature point set, wherein the target local feature point pair comprises a first feature point in the source image and a second feature point in the target image;
determining a mapped point of the first feature point on the target image;
determining a distance between the second feature point and the mapped point;
acquiring a quantity of reference local feature point pairs, wherein the reference local feature point pairs are target local feature point pairs with distances being less than a target distance threshold;
determining the matched image pair as a target image pair based on the quantity being greater than a target quantity;
determining a repetition rate of the target video and the source video based on the quantity of the target image pairs and the quantity of images of the target video; and
determining, based on the repetition rate greater than a target value, that the target video is as a video acquired by copying the source video.

5. The method according to claim 4 , wherein said acquiring the matched image pairs between the source video and the target video comprises:
acquiring image features of the images in the source video and the target video;
determining, based on the image features of the images, a similarity between the image belonging to the source video and the image belonging to the target video; and
determining the two images meeting a similarity requirement as a matched image pair.

6. The method according to claim 5 , wherein before said acquiring the image features of the images in the source video and the target video, the method further comprises:
extracting, based on an image feature extraction model, the image features of each of the images in the source video and the target video; and
storing the image features of the each of the images.

7. The method according to claim 6 , wherein the image features comprise a global image feature with a dimension being less than a target dimension value.

8. The method according to claim 5 , wherein the image features comprise a global image feature with a dimension being less than a target dimension value.

9. An electronic device, comprising:
a processor; and
one or more memories for storing at least one instruction executable by the processor;
wherein the at least one instruction, when executed by the processor, causes the processor to perform a method comprising:
acquiring a first feature point set of a source image and a second feature point set of a target image;
determining a target local feature point pair based on the first feature point set and the second feature point set, wherein the target local feature point pair comprises a first feature point in the source image and a second feature point in the target image;
determining a mapped point of the first feature point on the target image;
determining a distance between the second feature point and the mapped point;
acquiring a quantity of reference local feature point pairs, wherein the reference local feature point pairs are target local feature point pairs with distances being less than a target distance threshold; and
determining, based on the quantity being greater than a target quantity, that the target image is an image acquired by copying the source image,
wherein said determining the mapped point of the first feature point on the target image comprises:
determining a rigid body transformation matrix based on coordinates of the first feature point in the source image and coordinates of the second feature point in the target image;
acquiring a homography matrix based on the rigid body transformation matrix; and
determining the mapped point based on the homography matrix, wherein the mapped point is formed by mapping the first feature point to the target image.

10. The electronic device according to claim 9 , wherein said determining the target feature point pair based on the first feature point set and the second feature point set comprises:
determining distances between any local feature point in the first feature point set and a plurality of reference feature points, wherein the reference feature points are local feature points in the second feature point set;
acquiring a first distance and a second distance, wherein the first distance is a distance between the local feature point and a closest reference feature point, and the second distance is a distance between the local feature poi

Assignees

• Beijing Dajia Internet Information Tech Co Ltd

Inventors

• Yuan Yong
• Li Yan
• Liu An
• Yang Jie
• Ke Bingqing
• Ma Yanjun
• Zhang Cunyi
• Ding Dong

Classifications

• G06V20/40

  in video content (extracting overlay text G06V20/62; video retrieval G06F16/70; processing of video elementary streams in video servers H04N21/234; processing of video elementary streams in video clients H04N21/44) · CPC title

• G06V20/46 Primary

  Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames · CPC title

• G06T7/0002 Primary

  Inspection of images, e.g. flaw detection · CPC title

• G06V10/757

  Matching configurations of points or features · CPC title

• G06V10/44

  Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components · CPC title