Thanks for the great work.
Looking at the experiments, I saw that various papers in the NeRF series have experimented with the KITTI-360.
As I am still a freshman, I would like to have the experience of doing reconstruction using KITTI-360's poses and images in NeRF series(like NeRF, Mip360, ...).
However, if I simply use the poses and images of the dataset to learn, the performance is not good (sadly, the test PSNR is about 15~17...).
Is there something more I should be considering?
And where can I get this information?
Thanks in advance for your kind response.
Thanks for the great work.
Looking at the experiments, I saw that various papers in the NeRF series have experimented with the KITTI-360.
As I am still a freshman, I would like to have the experience of doing reconstruction using KITTI-360's poses and images in NeRF series(like NeRF, Mip360, ...).
However, if I simply use the poses and images of the dataset to learn, the performance is not good (sadly, the test PSNR is about 15~17...).
Is there something more I should be considering?
And where can I get this information?
Thanks in advance for your kind response.