Robustness visual tracking method based on transductive effect

Abstract

The invention discloses a robustness visual tracking method based on a transductive effect. The method comprises the following steps that an initial target is manually obtained in the first frame; in the subsequent frames, a great number of candidate targets are obtained through particle filtering sampling and are used as unlabeled samples, the former tracking results are used as labeled samples, and all samples are used as the peaks for building a map; through the cost function minimization, the confidence degree that candidate targets belongs to real targets is obtained, and in addition, the candidate target with the highest confidence degree is used as the current estimation; finally, the sample set is updated, the samples with the high confidence degree are added to a positive sample set, and the samples with low confidence degree are deleted. The method provided by the invention has the advantages that the problem of great appearance change of the moving targets is solved, the visual target robustness tracking method based on map transductive learning is provided, the method can be widely applied to video monitoring in various public places, and can also be used for airborne reconnaissance and monitoring systems at the same time, the considerable military benefits can be obtained, and the public safety service can be improved.

Description

technical field [0001] The invention belongs to the technical field of computer vision and pattern recognition, and in particular relates to a robust visual tracking method based on transduction. Background technique [0002] As one of the most basic core technologies in the field of computer vision, moving target tracking is the basis for various subsequent advanced processing, such as target recognition, target behavior analysis, video image compression and coding, and other high-level video processing and application understanding. The essence of target tracking is to calculate the position, size and movement speed of the target in each frame of image by analyzing the video sequence captured by the image sensor. The difficulty is that the image is a projection from a three-dimensional space to a two-dimensional plane, and there is information loss in itself, and the moving target is not a definite signal, it will undergo various complex changes such as rotation, scaling, ...

AI Technical Summary

Problems solved by technology

However, for images that vary both internally and externally, it is difficult to find a suitable template to effectively describe the target

An improved method is to keep the template updated to adapt to the change of the target, but this will also cause problems: the accumulation of errors makes the template deviate from the real target

The disadvantage of this method is that they only focus on the target information and completely ignore the background information, but in practice, the background information is very helpful for target tracking

[0014] 2. Graph Embedded Tracking Model Technical Solution and Its Defects

[0022] However, this method has two underlying assumptions: (1) the target is a Gaussian distribution; (2) the target and the background are linearly separable

In addition, the selection of negative samples in this method will directly affect the tracking results

[0023] However, the existing technology does not have a good method for dealing with large changes in target features, and its use is limited

Images