/SlaBins

Official repository for SlaBins: Fisheye Depth Estimation using Slanted Bins on Road Environments (ICCV 2023)

Primary LanguagePythonMIT LicenseMIT

SlaBins

This is the official repository of ICCV 2023 paper "SlaBins: Fisheye Depth Estimation using Slanted Bins on Road Environments". Paper link | Project page | Datasets

Methodology

example input output

To train our model, we pre-made camera lookup tables about SynWoodScape and KITTI-360 dataset using same code in OmniDet. We will provide our LUTs and the data preprocessing codes.

Unfortunately, model codes are not available because the work was corporated with company.

Datasets

We trained and evaluated our method on two fisheye datasets SynWoodScape, and KITTI-360. Because of the lack of images on SynWoodScape dataset (only 500 sequences are pre-released) and fixed camera slanted angle on KITTI-360 dataset, we used both datasets with our angle augmentation.

Our augmentation codes are available in this repository, and augmented datasets could be downloaded in the GoogleDrive.

Citation

If you found our code helpful for your research, please cite our paper as:

 @InProceedings{Lee_2023_ICCV,
    author    = {Lee, Jongsung and Cho, Gyeongsu and Park, Jeongin and Kim, Kyongjun and Lee, Seongoh and Kim, Jung-Hee and Jeong, Seong-Gyun and Joo, Kyungdon},
    title     = {SlaBins: Fisheye Depth Estimation using Slanted Bins on Road Environments},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2023},
    pages     = {8765-8774}
}