Method and device for evaluating stability of visual tracking algorithm based on intertemporal overlap rate

Abstract

The invention relates to the field of computer vision, in particular to a cross-time overlapping ratio-based visual tracking algorithm stability assessment method and device, and aims at solving the problem that the stability assessment of visual tracking algorithms is incorrect. The method specifically comprises the following steps of: aligning positions, in all the frames, of a same object in avideo to a same position; normalizing a tracking result of a certain visual tracking algorithm according to same alignment parameters; in a time span, calculating overlapping ratios between all the normalized tracking results and a normalized object labeling box; averaging the cross-time overlapping ratio on all the frames of the whole video so as to obtain a cross-time overlapping ratio score under the video; and averaging the cross-time overlapping ratios on all the videos and all the object categories of a certain object, and carrying out averaging on the basis of the time span so as to obtain a stability assessment result, under a complete visual tracking data set, of the visual tracking algorithm. The method and device are capable of effectively reflecting the tracking stability and correctness of visual tracking algorithms.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a method and device for evaluating the stability of a visual tracking algorithm based on a cross-temporal overlap rate. Background technique [0002] In recent years, fields such as computer vision, artificial intelligence, and machine perception have developed rapidly. Visual object tracking, as a fundamental problem in computer vision, has also been greatly developed. Object tracking is the process of using computer algorithms to process and analyze the images of consecutive frames in the video, so as to automatically track the position of the object of interest in the video, and obtain a stable trajectory of the object over a period of time. Traditional algorithms build complex tracking algorithms by building complex motion models, appearance models, memory models, and feature expressions. In recent years, with the widespread use of deep learning, many simple and effective trac...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above-mentioned problems in the prior art, that is, in order to solve the problem of inaccurate evaluation of the stability evaluation of the visual tracking algorithm, the first aspect of the present invention proposes a method for evaluating the stability of the visual tracking algorithm based on the cross-temporal overlap rate, Include the following steps:

Images