How should I generate the "landmarks/landmarks-1000v10" and "landmarks/visibility-1000v10_depth_normal" files for my own dataset? #22
Description
Hello, thank you for open-sourcing this excellent project. I have a question now. How can I obtain the "landmarks/landmarks-1000v10" and "landmarks/visibility-1000v10_depth_normal" files for each scene in the indoor-6 dataset in my own dataset? Currently, in my dataset, I only have images, SFM point cloud scenes generated from the images, 6DOF poses for each image viewpoint, and camera intrinsics. How should I generate these two files in my own dataset in order to proceed with the training?