Open shiyue-code opened 3 months ago
I am very confused by this line of code. union = pred_segms[pred_label]['area'] + gt_segms[gt_label]['area'] - intersection - gt_pred_map.get((VOID, pred_label), 0) Why subtract the GT background prediction errors?
union = pred_segms[pred_label]['area'] + gt_segms[gt_label]['area'] - intersection - gt_pred_map.get((VOID, pred_label), 0)
I am very confused by this line of code.
union = pred_segms[pred_label]['area'] + gt_segms[gt_label]['area'] - intersection - gt_pred_map.get((VOID, pred_label), 0)
Why subtract the GT background prediction errors?