Hi @YoYo000, I found that the MVSNet paper said that the ground truth point cloud is processed by screened Poisson surface reconstruction (SPSR) to generate the mesh surface, and then render the mesh to each viewpoint to generate the depth maps.
Could you please introduce more details about this process?
Currently, I am looking forward to generate some depth maps from a reconstructed point cloud as you did in the preprocessed datasets. Could you please give some suggestions? Thanks a lot.
Hi @YoYo000, I found that the MVSNet paper said that the ground truth point cloud is processed by screened Poisson surface reconstruction (SPSR) to generate the mesh surface, and then render the mesh to each viewpoint to generate the depth maps. Could you please introduce more details about this process? Currently, I am looking forward to generate some depth maps from a reconstructed point cloud as you did in the preprocessed datasets. Could you please give some suggestions? Thanks a lot.