PacGDC_large

PacGDC: Label-Efficient Generalizable Depth Completion with Projection Ambiguity and Consistency

Haotian Wang, Aoran Xiao, Xiaoqin Zhang, Meng Yang, and Shijian Lu

International Conference on Computer Vision (ICCV), October 2025

📝 Abstract

PacGDC is a label-efficient technique that enhances data diversity with minimal annotation effort for generalizable depth completion. It builds on novel insights into inherent ambiguities and consistencies in object shapes and positions during 2D-to-3D projection, allowing the synthesis of numerous pseudo geometries for the same visual scene. This process greatly broadens available geometries by manipulating scene scales of the corresponding depth maps. To leverage this property, we propose a new data synthesis pipeline that uses multiple depth foundation models as scale manipulators. These models robustly provide pseudo depth labels with varied scene scales, affecting both local objects and global layouts, while ensuring projection consistency that supports generalization. To further diversify geometries, we incorporate interpolation and relocation strategies, as well as unlabeled images, extending the data coverage beyond the individual use of foundation models.

🚀 Quick Start

Please refer to our Github Repo

📚 Citation

If you find our work useful, please cite:

@article{wang2025pacgdc,
  title     = {PacGDC: Label-Efficient Generalizable Depth Completion with Projection Ambiguity and Consistency},
  author    = {Wang, Haotian and Xiao, Aoran and Zhang, Xiaoqin and Yang, Meng and Lu, Shijian},
  journal   = {arXiv preprint arXiv:2507.07374},
  year      = {2025},
  url       = {https://arxiv.org/abs/2507.07374}
}