Method and system for searching similar video clips, server and storage medium
A video clip and video technology, applied in the field of video processing, can solve problems such as powerlessness, achieve the effect of small construction and access costs, and save computing power
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0055] Please refer to figure 1 , the embodiment of the present invention provides a kind of method of finding similar video segment, is used for comparing the similarity of a plurality of comparative videos in the video library and the video library one by one, to find out in the video library similar to the video to be tested or repeated video clips. The method includes:
[0056] Step S1, pre-training the convolutional neural network system to recognize objects.
[0057] First, through the Python open source web crawler tool, millions of pictures and object names of various objects are obtained from the Internet to form an object recognition material library. Then, establish a multi-layer convolutional neural network system, and train the convolutional neural network system through the pictures and names in the material library, so that the trained convolutional neural network system can recognize thousands of objects in the picture, And can identify the object.
[0058]...
Embodiment 2
[0069] A way to find similar video clips, see figure 2 , on the basis of embodiment 1, also includes the following steps:
[0070] Step S6, performing gray-scale thumbnail calculation on each frame of similar continuous video clips to obtain multiple gray-scale thumbnail data; obtaining any frame of pre-calculated similar continuous video clips from the comparison video The first grayscale thumbnail data;
[0071] Step S7, judging whether the similar continuous video segments belong to repeated segments according to the first grayscale thumbnail data and multiple grayscale thumbnail data.
[0072] In this embodiment, the similarity verification in steps 6 and 7 above is performed on the basis of the similar continuous video segments found in embodiment 1. Specifically, it is further judged whether the found similar continuous video segments are repeated segments, that is, segments with the same content, through the grayscale thumbnail data of the pictures. The method furth...
Embodiment 3
[0085] A kind of system of searching similar video segment of the present invention, see Figure 6 , which is used to find similar video segments in the comparison video and the video to be tested, including:
[0086] Building unit 1 for pre-training the convolutional neural network system to recognize objects;
[0087] Extracting unit 2, for obtaining the first picture sequence that the video to be tested is formed by drawing frames;
[0088] The identification unit 3 is used to identify the object in the first picture sequence through the trained convolutional neural network system, so as to obtain the first feature code used to identify the object;
[0089] The second feature code used to obtain the second picture sequence formed by frame extraction in advance of the comparison video and then identify and identify it;
[0090] The first comparison unit 4 is configured to compare the first feature code with the second feature code to find similar continuous video segments....
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More