Target tracking method and device based on anchor-point-free twin network key point detection

Abstract

The invention provides a target tracking method and device based on anchor-point-free twin network key point detection. The method comprises the steps: acquiring a target tracking video; sending the target template image and the target search area image into a pre-trained central point estimation module to generate a central point position estimation heat map, an offset estimation heat map from the upper left corner to the central point and a central point position error estimation heat map; sending the target search area image to a pre-trained corner estimation module to generate a corner position estimation heat map and a corner position error estimation heat map; and estimating angular point coordinates according to the offset estimation heat map from the upper left corner to the central point, determining a target frame in the current frame image according to the central point and the angular point coordinates, and completing target tracking. Target tracking is regarded as a determination problem of a center point and a top left corner point, and is decomposed into two parts of corner point position estimation and center point position estimation, so that the use of preset anchor points is avoided, the output quantity of heat maps is reduced, the parameter quantity of a network is reduced, and the speed of a tracking algorithm is increased.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a target tracking method and device for generating anchor-free twin network corner points. Background technique [0002] Generally speaking, target tracking is based on the target object that needs to be tracked in the first frame of the video, and the area where the target is located in the subsequent video frames is determined through the continuous reasoning of the tracking method. Currently, some tracking methods use an anchor-free Siamese network to calculate the similarity between the image area of ​​the target to be determined and the target template image, and then determine the target template image in subsequent frames according to the maximum similarity. [0003] In the 2 papers entitled "Siamese Box Adaptive Network for Visual Tracking" and "SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking" published at the 2020 I...

AI Technical Summary

Problems solved by technology

[0007] In view of the above problems, the present invention provides a target tracking method and device based on anchor-free twin network key point detection, which effectively solves the technical problems of the existing target tracking methods such as low accuracy and large amount of computation

Images