Video target detection method based on local area search

Abstract

The invention discloses a video target detection method based on local area search, and the method sequentially comprises the following steps: firstly detecting a first frame of a video through a single-frame detector, and obtaining the areas and types of all targets in an image; mapping the position of each target in the previous frame to the current frame from the second frame, inputting the area into a motion prediction module for detection to obtain the motion direction of the target, and finely adjusting the position of the target to obtain the area of the target in the current frame andthe result that the category of the target follows the previous frame; inputting the region information of the target and the region information of the target of the previous frame into a similarity detection module at the same time, judging whether the two regions represent the same target or not, and if not, detecting the position and the category of the target of the current frame by using a single-frame detector; and continuing the process on the rest frames until the detection is finished. The method provided by the invention can ensure the detection accuracy, improve the speed and save the resource overhead.

Description

technical field [0001] The invention relates to a video target detection method, in particular to a video target detection method based on local area search. Background technique [0002] The task of video target detection is to identify the position of each target contained in each frame of image in the input video and the category of each target. The position of the target is generally represented by a rectangular bounding box surrounding the target. The purpose of video target detection is to train the machine to detect the location and classification of objects contained in the video, so as to replace the manual processing of the huge video library. With the continuous improvement of Internet bandwidth, the application scenarios of video are becoming more and more extensive. For example, video websites and short video application platforms are gradually becoming popular among ordinary users. As a result, a large number of unclassified videos are generated on the Internet...

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned defects in the prior art, the task of the present invention is to provide a video object detection method based on local area search, aiming at the problem of low detection efficiency caused by high similarity between adjacent video frames, saving resource overhead, and taking into account while ensuring efficiency Detection accuracy

Images