Open sjg02122 opened 5 months ago
Thank you for your interest.
Unfortunately, the codes are confidential. However, If you want to reproduce our work, we can give you some hints.
Our core codes are based on the previous bin-based monocular depth estimation framework, AdaBins, and we add some transformation functions(Euclidean domain
-> Slanted Cylinder domain
) using pre-computed fisheye lookup-tabels(LUTs). The LUTs are computed same manner as OmniDet.
Thank you for your answer.
I can understand the confidentiality of the model part. However, I understand that the part about the database used the public dataset. I would appreciate it if you could partially release the part about the dataset.
As written in the paper, because of the lack of the SynWoodScape dataset (partially released), we used the SynWoodScape and KITTI-360 datasets with our angle augmentation. The augmented datasets will be released on this repository within this month.
Thank you.
Thank you.
Hi, thanks for the great work and the contribution. Will the augmented datasets be released soon? Thanks.~
Thank you.
Hi, thanks for the great work and the contribution. Will the augmented datasets be released soon? Thanks.~
Nowadays, I try to clean up our pre-processing codes and collect our augmented datasets for release. They will be shared soon on this repository. Sorry for being late.
Thank you.
Thank you.
Hi, thanks for the great work and the contribution. Will the augmented datasets be released soon? Thanks.~
Nowadays, I try to clean up our pre-processing codes and collect our augmented datasets for release. They will be shared soon on this repository. Sorry for being late.
Thank you.
Hi there~ Sorry to bother. Will the augmentation code (or dataset) be released soon? Thank you.
Dear all, I am sorry for the late release of the dataset. I just uploaded the datasets and our angle-augmented datasets are now available at Download Link.
If there is any problem, please report it to me for quick debugging.
Thank you.
Hi, thanks for the great work. Is there any plan about the code publishing?