Method and apparatus for searching video segment, device, and medium

Patent metadata

Field	Value
Publication number	US-11625433-B2
Application number	US-202117182960-A
Country	US
Kind code	B2
Filing date	Feb 23, 2021
Priority date	Apr 9, 2020
Publication date	Apr 11, 2023
Grant date	Apr 11, 2023

Abstract

Embodiments of the present disclosure disclose a method and apparatus for searching a video segment, a device and a medium, and relate to the field of video data search. The method includes: sampling video frames from a target video and videos to be searched in a video library, and extracting features from the sampled frames; matching the target video and the videos to be searched according to the extracted features to determine a candidate video to be searched that matches the target video; determining at least one candidate video segment from the determined candidate video, and calculating a degree of matching between the target video and each candidate video segment based on the extracted features of each sampled frame; and determining a video segment matching the target video in the videos to be searched according to the calculated degree of matching between the target video and each candidate video segment.

First claim

What is claimed is:

1. A method for searching a video segment, comprising:
sampling video frames from a target video and videos to be searched in a video library, and extracting features from the sampled video frames;
matching the target video and the videos to be searched according to the extracted features to determine a candidate video to be searched that matches the target video;
determining at least one candidate video segment from the determined candidate video to be searched, wherein a first ratio of a length of a first candidate video segment alone to a length of the target video is within a preset interval, and a second ratio of a length of a second candidate video segment alone to the length of the target video is within the preset interval, and calculating a degree of matching between the target video and each candidate video segment based on the extracted features of each sampled frame; and
determining a video segment matching the target video in the videos to be searched according to the degree of matching between the target video and each candidate video segment.

2. The method according to claim 1 , wherein the matching comprises:
determining, according to the extracted features of the sampled video frames, matching results between the sampled video frames in the target video and the sampled video frames in the videos to be searched; and
counting the matching results, and determining, according to the counted matching results, the candidate video to be searched that matches the target video.

3. The method according to claim 1 , wherein the determining at least one candidate video segment from the candidate video to be searched comprises:
combining every two sampled video frames of the candidate video to be searched, to obtain at least one sampled frame pair, and using the two sampled video frames in the sampled frame pair as a start frame and an end frame of a candidate video segment to obtain the at least one candidate video segment.

4. The method according to claim 1 , wherein the calculating comprises:
calculating a similarity between each sampled frame in the target video and each sampled frame in the candidate video segment based on the extracted features of each sampled frame; and
determining the degree of matching between the target video and the candidate video segment based on the similarity between each sampled frame in the target video and each sampled frame in the candidate video segment.

5. The method according to claim 4 , wherein the determining the degree of matching between the target video and the candidate video segment comprises:
determining a similarity between the target video and the candidate video segment based on the similarity between each sampled frame in the target video and each sampled frame in the candidate video segment; and
superimposing the similarity between the target video and the candidate video segment with a preset value of a reward function to obtain the degree of matching between the target video and the candidate video segment,
wherein the preset value of the reward function is positively correlated with a length of the candidate video segment.

6. The method according to claim 1 , wherein the determining a video segment matching the target video in the videos to be searched comprises:
determining, from the candidate video segment, a target candidate video segment whose degree of matching with the target video satisfies a preset condition; and
searching previous frames based on a position of a sampled frame serving as a start frame of the target candidate video segment in a corresponding video to be searched, searching subsequent frames based on a position of the sampled frame serving as an end frame of the target candidate video segment in the corresponding video to be searched, and determining a start frame and an end frame of the video segment matching the target video in the corresponding video to be searched.

7. An electronic device, comprising:
one or more processors; and
a storage apparatus, storing one or more programs,
wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform an operation for searching a video segment, comprising:
sampling video frames from a target video and videos to be searched in a video library, and extracting features from the sampled video frames;
matching the target video and the videos to be searched according to the extracted features to determine a candidate video to be searched that matches the target video;
determining at least one candidate video segment from the determined candidate video to be searched, wherein a first ratio of a length of a first candidate video segment alone to a length of the target video is within a preset interval, and a second ratio of a length of a second candidate video segment alone to the length of the target video is within the preset interval, and calculating a degree of matching between the target video and each candidate video segment based on the extracted features of each sampled frame; and
determining a video segment matching the target video in the videos to be searched according to the degree of matching between the target video and each candidate video segment.

8. The electronic device according to claim 7 , wherein the matching comprises:
determining, according to the extracted features of the sampled video frames, matching results between the sampled video frames in the target video and the sampled video frames in the videos to be searched; and
counting the matching results, and determining, according to the counted matching results, the candidate video to be searched that matches the target video.

9. The electronic device according to claim 7 , wherein the determining at least one candidate video segment from the candidate video to be searched comprises:
combining every two sampled video frames of the candidate video to be searched, to obtain at least one sampled frame pair, and using the two sampled video frames in the sampled frame pair as a start frame and an end frame of a candidate video segment to obtain the at least one candidate video segment.

10. The electronic device according to claim 7 , wherein the calculating comprises:
calculating a similarity between each sampled frame in the target video and each sampled frame in the candidate video segment based on the extracted features of each sampled frame; and
determining the degree of matching between the target video and the candidate video segment based on the similarity between each sampled frame in the target video and each sampled frame in the candidate video segment.

11. The electronic device according to claim 10 , wherein the determining the degree of matching between the target video and the candidate video segment comprises:
determining a similarity between the target video and the candidate video segment based on the similarity between each sampled frame in the target video and each sampled frame in the candidate video segment; and
superimposing the similarity between the target video and the candidate video segment with a preset value of a reward function to obtain the degree of matching between the target video and the candidate video segment,
wherein the preset value of the reward function is positively correlated with a length of the candidate video segment.

12. The electronic device according to claim 7 , wherein the determining a video segment matching the target video in the videos to be searched comprises:
determining, from the candidate video segment, a target candidate video segment whose degree of matching with the target video satisfies a preset condition; and

Assignees

• Beijing Baidu Netcom Sci & Tech Co Ltd

Inventors

• Long Xiang
• Wang Ping
• Li Fu
• He Dongliang
• Sun Hao
• Wen Shilei

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06V20/48

  Matching video sequences · CPC title

• G06V20/46

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

• G06F16/7847 Primary

  using low-level visual features of the video content · CPC title

• G06N3/08

  Learning methods · CPC title