Open Parskatt opened 1 week ago
Thanks for the suggestion. For now, we do have a file that stores reconstruction metadata for the web-viewer. It can be found here. It is a dictionary with format:
{
"cat_id": [
"cat_name",
num_recons,
sparse_0_num_imgs,
sparse_0_num_points3D,
sparse_1_num_imgs,
sparse_1_num_points3D,
...
],
...
}
It would be nice to have some reconstruction metadata about the scenes without having to download the full reconstructions. For example, no. cameras, no. 3D points.