Hi @shurans, I was going through perpareNYUCADdata.m to try to generate semantic labels for NYUCAD. The line where we convert the meshes to .obj files are commented out and the function file is missing.
From the codes, my guess is that the pipeline to generate the labels are:
- Gather all the faces, vertices, and class ID in meshobjArr.
- Convert these into a single .obj file.
- renderView takes in the .obj file and the camPose to return a depth map and the semantic label image.
I am wondering how is the class ID being stored in the .obj file so that renderView can know what label to return as a label.png? Alternatively, may I ask for the mesh2obj.m file so I can study it further?
Thank you for your time.
Hi @shurans, I was going through
perpareNYUCADdata.mto try to generate semantic labels for NYUCAD. The line where we convert the meshes to .obj files are commented out and the function file is missing.From the codes, my guess is that the pipeline to generate the labels are:
I am wondering how is the class ID being stored in the .obj file so that
renderViewcan know what label to return as alabel.png? Alternatively, may I ask for themesh2obj.mfile so I can study it further?Thank you for your time.