The evaluation script uses IoU 0.5 for association and 0.9 score for thresholding both methods. We can see that our final output achieves 80.9% in MOTA, 7.8% better than Hungarian, as well as 20% better on MT, 10% lower on ML, while still having similar MOTP. Motion Forecasting: We ...
The evaluation script uses IoU 0.5 for association and 0.9 score for thresholding both methods. We can see that our final output achieves 80.9% in MOTA, 7.8% better than Hungarian, as well as 20% better on MT, 10% lower on ML, while still having similar MOTP. Motion Forecasting: We ...