Project Icon

mahalanobis_3d_multi_object_tracking

Improving Accuracy in Autonomous 3D Multi-Object Tracking Using Probabilistic Techniques

Product DescriptionThe project presents a probabilistic approach to 3D multi-object tracking aimed at enhancing accuracy in autonomous driving systems. By integrating MEGVII detection inputs, the method surpasses the AB3DMOT baseline, earning first place in the NuScenes Tracking Challenge. It features a combination of Kalman Filter covariance estimation and data association techniques, improving tracking accuracy, particularly for small objects such as pedestrians. Open-source code and setup guidelines are available for developers interested in replicating or further exploring these results.
Project Details