Target tracking method and device, electronic equipment and computer readable storage medium

Abstract

The invention discloses a target tracking method applied to the technical field of target tracking, and comprises the following steps: obtaining a search image based on a current frame image, and obtaining a confidence value and a coordinate box according to a network tracking model; selecting a first target number of candidate boxes according to the confidence value and the coordinate box; obtaining a tracking result according to the historical tracking value and the candidate box; obtaining a corresponding confidence value, and judging whether the confidence value is within a first preset range or not; if so, performing image cutting on the current frame image according to the target template image; and under the condition that the definition of the cropped image is not less than the preset definition, replacing the cropped image corresponding to the current video frame image with a new target template image. By applying the embodiment of the invention, when the template is updated,whether to update or not is judged by combining the result fed back by the tracking model and the quality of the target image in the current frame, so that the precision of the tracking result is improved, and the complex calculation amount is avoided.

Description

technical field [0001] The present invention relates to the technical field of video target tracking, in particular to a target tracking method, device, electronic equipment and computer-readable storage medium. Background technique [0002] With the development of science and technology, convolutional neural network method (CNN) has played an increasingly important role in image and video processing. The most well-known face recognition technology has been very mature, and many application scenarios have brought convenience to people. Life enjoyment, such as face sign-in, mobile phone face unlocking, face check-in and face payment, etc. [0003] The tracking target in the video will change in different time periods due to various factors (such as camera shooting, target movement, shooting scene, etc.). However, in the prior art, a fixed target template is usually used for target tracking, which will increase the difficulty of tracking model judgment and reduce the accuracy...

AI Technical Summary

Problems solved by technology

However, in the prior art, a fixed target template is usually used for target tracking, which will increase the difficulty of tracking model judgment and reduce the accuracy of tracking results when the target template is fixed and changes occur. , if a target template is set for each video frame image, it will cause an increase in the amount of calculation

Therefore, there is a lack of an effective method for updating the target template in the prior art

Images