Online learning offline video tracking method based on key frames

Abstract

The invention discloses an online learning offline video tracking method based on key frames. The method includes the following steps that the key frames in a certain number are selected to label a given offline video, a complete template dictionary is established, and in the tracking process, the distance between each candidate image region and pure template sub-blocks is calculated, so that the pure template sub-blocks are effectively selected, and calculation efficiency is improved; in order to reduce accumulation of tracking errors, an open loop problem is converted into a closed loop problem through a cyclic tracking strategy, and therefore the target tracking robustness is effectively improved; through the tracking result, a dynamic template is online learnt in a semi-supervised mode so as to adapt to the change of the target appearance. The target template dictionary is effectively learnt online in the tracking process, so that error accumulation caused by template updating each time is avoided; the open loop problem is converted into the closed loop problem through the cyclic tracking strategy, so that the target tracking robustness is effectively improved.

Description

technical field [0001] The invention relates to the technical field of computer vision tracking, in particular to an offline video tracking method based on key frame online learning. Background technique [0002] According to different video sources, video object tracking can be simply divided into two categories: online video object tracking and offline video object tracking. The object tracking of online video only has the video data before the current frame, so the object tracking problem of online video is an open-loop control system, which makes the error inevitably accumulate. The target tracking of offline video already has a complete video before tracking, so a small number of key frames can be marked to transform the open-loop control system into a closed-loop control system, so that offline video tracking can be used for video labeling and video retrieval. , event analysis, and video compression based on moving objects. [0003] In general, the current target tra...

AI Technical Summary

Problems solved by technology

The target tracking of online video only has the video data before the current frame, so the target tracking problem of online video is an open-loop control system, which makes the error inevitably accumulate

[0003] In general, the current target tracking algorithm mainly has two key issues: (1) appearance model; (2) tracking framework

Then, the robust estimation strategy, Yang's data-driven enhanced adaptive method, Liao's robust Kalman filter-based tracking method, and Gai and Stevenson's method based on dynamic models, although in some specific scenarios obtained better Good tracking performance, but there are certain shortcomings: that is, all the above-mentioned subspace-based tracking algorithms must first expand the image into a one-dimensional vector, and the spatial distribution information of the target's appearance is almost completely lost, so that the model has a good understanding of the target's apparent Very sensitive to global changes and noise

[0006] Although the appearance model based on sparse representation has achieved great success in dealing with occlusion and noise, the model still has the following problems: the number of target templates in the template dictionary is too small (usually 10), which is far from the sparse representation theory. Over-complete requirements for dictionary templates

Such a dictionary update method may also lead to error accumulation. If the update frequency is too fast, the error accumulation will be large.

On the other hand, if the update is too slow, it will be difficult to adapt to changes in the target appearance

Images