Video information retrieval method and device based on RGB classification and storage medium

Abstract

The invention discloses a video information retrieval method and device based on RGB classification and a storage medium. The method comprises the following steps: quickly retrieving a video; based onthe advantages of a SqueezeNet small model structure, decomposing a video into a frame of picture by using FFMPEG software; respectively extracting RGB information of the frame pictures by using MATLAB software; classifying the three kinds of color information; inputting the RGB information of the classified picture into a SqueezeNet network model; extracting RGB effective information of the picture, inputting the information of the to-be-retrieved picture into the SqueezeNet neural network model as well; obtaining the RGB classification information of the to-be-retrieved picture; and comparing the RGB classification information of the to-be-retrieved picture with the RGB effective information of the video frame picture, judging which kind of RGB class information the to-be-retrieved picture belongs to, and searching a corresponding frame picture from the corresponding classified RGB class information, so that the corresponding video information at the time point can be directly found. According to the method, the image video retrieval calculation amount is greatly reduced, the video reading time is shortened, the working efficiency is improved, and the accuracy is greatly improved.

Description

technical field [0001] The invention belongs to the field of video technology, and in particular relates to a video information retrieval method, device and storage medium based on RGB classification. Background technique [0002] Image and video information retrieval is to use the distinguishing information in the image and video to retrieve the video. When the user needs to view specific information in the video, he needs to view the video from the beginning to the end, extract the texture, image, color and other information in the video, and record each The information of a frame of video, classify different information, and use these information to retrieve the information in the video, compare and judge whether it is the information you need, which greatly wastes time, because during the viewing process, The information of each frame is similar, and it is very likely that judgment errors will occur. In the process, the amount of calculation is large and the algorithm is...

AI Technical Summary

Problems solved by technology

[0002] Image and video information retrieval is to use the distinguishing information in the image and video to retrieve the video. When the user needs to view specific information in the video, he needs to view the video from the beginning to the end, extract the texture, image, color and other information in the video, and record each The information of a frame of video, classify different information, and use these information to retrieve the information in the video, compare and judge whether it is the information you need, which greatly wastes time, because during the viewing process, The information of each frame is similar, and it is very likely that judgment errors will occur. In this process, the amount of calculation is large and the algorithm is complicated, which has become a problem that needs to be improved urgently.

Images