Target tracking method based on space-time feature fusion learning

Abstract

The invention discloses a target tracking method based on spatio-temporal feature fusion learning, and relates to the technical field of computer vision and pattern recognition. The method comprises the steps that firstly, a space-time feature fusion learning network is constructed, space-time features comprise time sequence features and space features, and the time sequence features are extractedin the mode that Alexnet and a time recurrent neural network are combined; Wherein the spatial features are divided into target object spatial transformation features and background spatial features,and YOLOv3 and Alexnet extraction is adopted respectively. In the initial training process of the network, a training data set and a random gradient descent method are used for training the space-time feature fusion learning network, and after training is completed, the network can obtain the initial capacity for positioning the target object. The image sequence to be tracked is input into the network for forward processing, the network outputs the position and confidence of the target object bounding box, the confidence decides whether the network performs online learning or not, and the position of the bounding box realizes positioning of the target object, so that tracking of the target object is realized.

Description

technical field [0001] The invention relates to the technical fields of computer vision and pattern recognition. Background technique [0002] Visual object tracking is an important research topic in the field of computer vision. Its research content is to automatically identify the target object to be tracked in the subsequent video sequence according to a given video clip, and obtain the continuous position, appearance and motion information of the target. Target tracking is widely used in military and civilian intelligent monitoring, human-computer interaction, automatic control systems and other fields, and has strong practical value. But in reality, the appearance of target objects is easily affected by factors such as deformation, occlusion, and illumination changes, making visual object tracking a very challenging problem. At present, target tracking methods mainly include classical target tracking methods and deep learning target tracking methods. [0003] Classica...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a target tracking method based on spatio-temporal feature fusion learning, which can effectively solve the problem of tracking difficulty or even loss when the target object is deformed or occluded, and realize long-term real-time and accurate target tracking

Images