OpenCV 4 with Python Blueprints - Second Edition

Associating detections with trackers

In the Sort algorithm, decisions about whether two bounding boxes should be considered to be of the same object are made based on Intersection over Union. Previously in this chapter, you learned about this metric and implemented a function to compute it. Here, we'll define a function that will associate detection and tracking boxes based on their IoU value:

def associate_detections_to_trackers(detections: np.ndarray, trackers: np.ndarray,          iou_threshold: float = 0.3) -> Tuple[np.ndarray, np.ndarray, np.ndarray]:

The function accepts the bounding boxes of detections and the predicted boxes of trackers, as well as an IoU threshold. It returns matches as an array of pairs of corresponding indexes in the corresponding ...