vis_height is the resolution of output frame

:param gt: groundtruth entities for a frame :param pred: prediction entities for a frame :param iou_thresh:

The solver is to merge predictions from detection branch as well as from track branch. The goal is to assign an unique track id to bo

Arguments: x (list[Tensor]): feature maps for each level boxes (list[BoxList]): boxes to be used to perform the pooli

forward functions of the tracker :param features: raw FPN feature maps from feature backbone :param boxes: template bounding

Arguments: features (list[Tensor]): feature-maps from possibly several levels proposals (list[BoxList]): proposal box

Arguments: x (tuple[tensor, tensor]): x contains the class logits and the box_regression from the model.

Arguments: x (list[Tensor]): feature maps for each feature level. Returns: results (tuple[Tensor]): feature m

Kill a track

group conv2d to calculate cross correlation, fast version

for loop to calculate cross correlation, slow version