Dual-core KCF target tracking method based on space-time significance

Abstract

The invention provides a dual-core KCF target tracking method based on space-time saliency, and the method comprises the following steps: S1, extracting a target region, and extracting a salient region through a visual saliency model; s2, HOG features of a target frame and a significant frame are extracted respectively to train parameters of a filter; s3, respectively calculating response distribution maps yk and ys of the filters of the target frame and the significant frame based on the classification of the ridge regression classifier, wherein the coordinate positions corresponding to max (yk) and max (ys are the position of the target frame and the position of the salient frame; S4, calculating the offset of the target frame and the offset of the salient frame based on the position coordinates of the current frame and the previous frame, and carrying out weighted averaging on the offset to obtain a new offset value as the correction offset of the target frame; and S5, obtaining thecorrected target position of the current frame through the position of the previous frame and the correction offset of the target frame. The method is based on the KCF algorithm principle with good speed and accuracy at the present stage, and effectively solves the problem of target tracking drift.

Description

technical field [0001] The invention relates to the field of target tracking, in particular to a dual-core KCF target tracking method based on spatio-temporal saliency. Background technique [0002] Object tracking is one of the most active research fields in computer vision, and it is widely used in motion analysis, behavior recognition, monitoring, and human-computer interaction. The current research on target tracking technology has made great progress, and many tracking algorithms have emerged. The current mainstream tracking algorithms are mainly divided into two categories: [0003] One is the tracking algorithm based on deep learning, which is mainly based on the deep neural network framework for learning. Deep neural network has been applied in object tracking because of its powerful learning function in image feature extraction. For example, the target tracking method based on the fully convolutional neural network not only uses CNN as a tool for feature extracti...

AI Technical Summary

Problems solved by technology

However, the KCF algorithm uses linear interpolation to update the target model, resulting in the accumulation of target apparent information deviations in the tracking process, resulting in target drift, which easily leads to a decrease in target tracking accuracy in subsequent frames

In addition, the KCF algorithm uses a single HOG feature. Although it can capture the outline of the target well, it is easy to cause the target drift tracking to fail when the target is occluded.

Images