Visual tracking method based on semi-supervised features and filter combination learning

Abstract

The present invention discloses a visual tracking method based on semi-supervised features and filter combination learning. The method comprises the steps of: extracting training samples according totarget position of a (t-1)th frame, and extracting direction gradient histogram features, gray features and color features; taking features of the pixel points as nodes, constructing a 8 neighbourhoodmap by taking two pixel points belonging to the same class of probability value as an edge weight, and calculating an initial weight vector according to the positions where the pixel points are located; and finally, constructing a model to jointly solve a filter and the weight vectors, setting a search area at the target position of the (t-1)th frame and extracting features on the (t-1)th image,and employing the weight vectors to perform weighing of the features and finally perform convolution with the filter to obtain a response map and determine the center point of the target. The visual tracking method employs the semi-supervised mode to perform combination learning of feature reliability and related filters in a uniform optimization frame to inhibit the interference of a background area on tracking in the tracking process and allow a trackers to have a better robust effect for the tracking target.

Description

technical field [0001] The invention relates to a visual tracking method, in particular to a visual tracking method based on joint learning of semi-supervised features and filters. Background technique [0002] Visual tracking is one of the important research topics in the field of computer vision. The definition of video tracking is: given the state information of the target in the initial frame of the video sequence, including the target position and size, and then predicting the process of the target motion state in the next video sequence. Visual tracking has a very wide range of applications in video surveillance, human-computer interaction, robotics and other fields. [0003] In recent years, the visual tracking algorithm based on correlation filtering has received widespread attention due to its excellent performance in accuracy and efficiency. detection. The method itself can effectively realize target positioning, but it requires a large number of training sample...

AI Technical Summary

Problems solved by technology

Because the correlation filter regards the entire picture as a positive sample and the sample obtained by circular shift is a negative sample, the learned filter is easily disturbed by background information.

Images