Video verification method, device, apparatus, and readable storage medium

A video and video frame technology, applied in the field of computer vision, can solve the problems of difficulty in ensuring accuracy and high cost of manpower review, and achieve the effect of improving accuracy and saving labor costs

Pending Publication Date: 2019-03-01
BAIDU ONLINE NETWORK TECH (BEIJIBG) CO LTD
View PDF3 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, in order to filter out these illegal videos, enterprises often spend a lot of manpower on reviewing. However, with the increasing number of illegal videos, it is difficult to rely solely on manual review, and the cost of manual review is high and the accuracy is difficult to guarantee.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Video verification method, device, apparatus, and readable storage medium
  • Video verification method, device, apparatus, and readable storage medium
  • Video verification method, device, apparatus, and readable storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0031] figure 1 It is a flow chart of a video verification method provided in Embodiment 1 of the present invention. This embodiment is applicable to the situation where a video to be verified is subjected to violation verification. The method can be executed by a video verification device, which can be implemented by hardware and / or Software constitutes and is generally integrated in electronic equipment, specifically including the following operations:

[0032] S110. Acquire a sequence of video frames in the video to be verified.

[0033] The sequence of video frames includes a plurality of video frames with consecutive time stamps. The number of video frame sequences can be one, two or more.

[0034] S120. Input the video frame sequence into the video verification model to obtain the confidence degree corresponding to the video frame sequence.

[0035] Optionally, the video frame sequence is sequentially input into the video verification model according to the inter-fram...

Embodiment 2

[0043] In this embodiment, on the basis of the foregoing embodiments, the video verification model is further refined. Further, the temporal feature extraction unit includes at least one one-dimensional convolution kernel, which is used to perform one-dimensional convolution on each spatial feature in the time domain to obtain a one-dimensional feature vector. Further, the confidence calculation unit includes at least one fully connected layer, which is used to fuse the spatio-temporal features to obtain the confidence. Further, the confidence calculation unit also includes a normalization layer, which is used to normalize the fused features to obtain the confidence. Further, the confidence calculation unit includes at least one fully connected layer and a classification layer; at least one fully connected layer is used to fuse spatio-temporal features, and the classification layer is used to classify the fused features to obtain confidence.

[0044] Figure 2aIt is a schema...

Embodiment 3

[0051] image 3 It is a flowchart of a method for determining a cover provided in Embodiment 3 of the present invention. The embodiment of the present invention adds operations on the basis of the technical solutions of the above-mentioned embodiments.

[0052] Further, before the operation "input the video frame sequence into the video verification model to obtain the confidence corresponding to the video frame sequence", the additional operation "obtain multiple sample video frame sequences; respectively obtain the labels corresponding to each sample video frame sequence, label Including compliance labels and violation labels; according to multiple sample video frame sequences and labels corresponding to each sample video frame sequence, the video verification model to be trained is trained" to pre-train the video verification model.

[0053] Such as image 3 A video verification method shown includes:

[0054] S310. Acquire a sequence of video frames in the video to be v...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The embodiment of the invention discloses a video verification method, an apparatus, a device and a readable storage medium. The method comprises the following steps: obtaining a video frame sequencein a video to be verified; inputting the video frame sequence into a video verification model to obtain a confidence level corresponding to the video frame sequence; if the confidence level meets thepreset requirements, verifying the video compliance to be verified; the video verification model comprises: A spatial violation feature extraction unit a time domain feature extraction unit and a confidence calculation unit. The spatial violation feature extracting unit is used for extracting the violation features of each video frame in the video frame sequence to obtain each spatial feature, thetime domain feature extracting unit is used for time convolution of each spatial feature to obtain the spatio-temporal feature, and the confidence degree calculating unit is used for calculating theconfidence degree of the spatio-temporal feature. The embodiment of the invention can improve the efficiency and the accuracy of the violation video verification.

Description

technical field [0001] The embodiments of the present invention relate to computer vision technology, and in particular to a video verification method, device, equipment and readable storage medium. Background technique [0002] In the era of mobile Internet, people have more and more ways to transmit and communicate information, from text-based to gradually relying more on various images and videos. However, while science and technology bring convenience to people, the massive content generated based on the Internet every day is full of a large amount of illegal information such as violence and terrorism. [0003] At present, in order to filter out these illegal videos, enterprises often spend a lot of manpower on reviewing. However, with the increasing number of illegal videos, it is difficult to rely solely on manual review, and the cost of manual review is high, and the accuracy is difficult to guarantee. Contents of the invention [0004] Embodiments of the present i...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/00H04N21/234H04N21/44
CPCH04N21/23418H04N21/44008G06V20/49
Inventor 赵翔刘霄文石磊李旭斌丁二锐孙昊李鑫柏提杨凡
Owner BAIDU ONLINE NETWORK TECH (BEIJIBG) CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap