Abstract
Aiming at the current stage of the twin network target tracking algorithm, the tracking target is occluded, the tracking is affected by illumination, and the target's scale change from far to near or from near to far causes tracking failure. This article will optimize and improve from two directions. The twin neural network first uses an adaptive detailed feature extraction, adds a residual network to the twin network, and embeds a detailed feature retention module in each layer, amplifies the changes in the target feature, and retains the important structure of the original target feature Details: Secondly, the introduction of a spatial attention mechanism allows the main branch to pay more attention to the area to be matched, improves the ability to distinguish features, and makes the tracking effect better. In order to verify the effectiveness of this experiment, this experiment was tested on the data set OTB2015. The experiment proved that the proposed algorithm performs better in accuracy and success rate.
Highlights
This paper introduces the spatial attention mechanism, which is to give different weights to different positions in different spatial positions during feature extraction, so as to achieve the extraction of different feature positions on the feature map, taking the same position between the feature maps as The unit, by comparing the similarity between the feature maps, pays attention to the importance of a certain position in the spatial position
This article judges the comprehensive performance of each algorithm based on the accuracy and success rate of target tracking
(2) As shown in formula (3): the calculation of the target tracking success rate refers to the area of the overlap rate (IoU) curve of the target's prediction block diagram and the true boundary block diagram
Summary
The SiamFC network is mainly used to calculate the similarity of the features extracted by the two branches in the twin network, and estimate the position of the target through the highest point of the similarity response. If the target image z and the candidate image x are the same target, the mapping function will return a higher similarity score, on the contrary, it will return a lower similarity score. To find the new position of the target in the new video frame image, we have to perform a complete test on all possible positions and select the candidate target with the highest similarity. The mapping function f(z, x) is shown in the following formula 1, where b_1 represents the offset signal of the position fetching point.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have