Trace template self-adaption method based on variance estimation

Abstract

The invention provides a trace template self-adaption method based on variance estimation. The method comprises the following steps that firstly, an initial center position of a target and an initial position frame determined according to the initial center position are acquired, wherein the initial center position is marked as x*; secondly, features of the target are acquired based on the initial position frame; thirdly, a target trace template is established according to the features acquired in the second step; fourthly, the target is traced according to the target trace template acquired in the third step, and the current time center position of the target is acquired; fifthly, the dimension of the target trace template is updated according to the current time center position of the target. By means of the method, trace robustness, timeliness and accuracy can be improved.

Description

technical field

[0001] The invention relates to the field of video target tracking, in particular to a tracking template adaptive method based on variance estimation. Background technique

[0002] Object tracking has always been one of the hot spots in the field of computer vision, and it is widely used in various fields such as motion analysis, behavior recognition, monitoring, and human-computer interaction. The difficulty of target tracking lies in the change of the target's own posture and appearance, the occlusion between targets, and the change of the environment affect the accuracy of tracking. The tracking algorithm based on online learning of appearance features has become one of the mainstream tracking algorithms in recent years because of its good performance. This type of algorithm regards target tracking as a binary classification problem of background and target, and distinguishes the target from the background through online learning of target templates, so a...

Images