Fig. 10
From: LeTra: a leaf tracking workflow based on convolutional neural networks and intersection over union
Flowchart illustrating the tracking process. Panel A depicts the initial step, where the mask for time step n (query) is retrieved and the IoU is computed for all masks at time step n + 1 (targets). In Panel B, the decision loop is made based on the resulting IoU. If the query and target match, the pair is assigned, and the target is removed from the target pool. If the query is not found, indicating that the leaf is either covered in the image or the detection failed, the position is stored by saving the same mask. Finally, the remaining target masks are considered as new detections. Panel C shows the list of masks, in order, which is stored and set as the new queries for the next time step