Curricular Object Manipulation in LiDAR-based Object Detection | IEEE Conference Publication | IEEE Xplore

Curricular Object Manipulation in LiDAR-based Object Detection


Abstract:

This paper explores the potential of curriculum learning in LiDAR-based 3D object detection by proposing a curricular object manipulation (COM) framework. The framework e...Show More

Abstract:

This paper explores the potential of curriculum learning in LiDAR-based 3D object detection by proposing a curricular object manipulation (COM) framework. The framework embeds the curricular training strategy into both the loss design and the augmentation process. For the loss design, we propose the COMLoss to dynamically predict object-level difficulties and emphasize objects of different difficulties based on training stages. On top of the widely-used augmentation technique called GT-Aug in Li-DAR detection tasks, we propose a novel COMAug strategy which first clusters objects in ground-truth database based on well-designed heuristics. Group-level difficulties rather than individual ones are then predicted and updated during training for stable results. Model performance and generalization capabilities can be improved by sampling and augmenting progressively more difficult objects into the training samples. Extensive experiments and ablation studies reveal the superior and generality of the proposed framework. The code is available at https://github.com/ZZY816/COM.
Date of Conference: 17-24 June 2023
Date Added to IEEE Xplore: 22 August 2023
ISBN Information:

ISSN Information:

Conference Location: Vancouver, BC, Canada

Contact IEEE to Subscribe

References

References is not available for this document.