在哪里用kalman预测了tracks的新位置？

[elvindp]

在论文里，提到用kalman预测了tracks的新位置，14-16行。 但在代码里，并没有用到预测的新位置。 # Predict the current location with KF STrack.multi_predict(strack_pool) dists = matching.iou_distance(strack_pool, detections) multi_predict用kalman预测了新的mean，covariance，但在iou_distance的计算里面，用的是旧的tlbr。

[ifzhang]

We use the updated mean of Kalman Filter here: https://github.com/ifzhang/ByteTrack/blob/191e983fa8d9a29173943cb75a771a90e527a47f/yolox/tracker/byte_tracker.py#L92

[elvindp]

We use the updated mean of Kalman Filter here:

https://github.com/ifzhang/ByteTrack/blob/191e983fa8d9a29173943cb75a771a90e527a47f/yolox/tracker/byte_tracker.py#L92

Becuase, it is after iou distance calculation, that the track will update, # Predict the current location with KF STrack.multi_predict(strack_pool) this prediction only use for next frame iou distance calculation. If move this part behind linear assignment, will not effect the results. Am I right?

[qikebai]

multi_predict更新了track的mean和covariant，mean的前四个数就是新box的位置 (center x, center y, ratio, height), 在iou_distance的计算里, 是把(center x, center y, ratio, height)转化成tlwh再转化成tlbr，然后再算iou