Algorithm 1 The optimized IG-SiamRPN bounding box regression algorithm.
Input:,,,,,
represents a set of detected bounding boxes in the form of , where and , respectively, represent the continuous coordinates of the upper left corner and the lower right corner of the area.
is the feature map of the grid area before PrRoI pooling.
is the number of iteration steps to fine-tune the bounding box, is the step size, is the threshold for stopping the iteration early, and is the tolerance of positioning degradation.
Output: predicted target bounding box after fine-tuning
