chaowang15 / plane-opt-rgbd Goto Github PK

Hi, this is a very great work!
I have a question about the mesh rendering in mesh_visibility. Camera intrinsic is not used in mesh rendering. Instead, a trans_camera matrix is generated with glm::lookAt function. I think the generated intrinsic matrix by glm::lookAt is not same as real camera intrinsic. This may induce some problems with the resulted visibility files. Is this right?

Hi, could you pls specify the compiler version (some other dependencies, if any), cause I've got so many errors during the process of compilation, for example:
'move' is not a member of 'std';
'unordered_map' was not declared in this scope;
...

Hi, thanks for the great work. I tried several images from BundleFusion dataset in VoxelHashing and got a good sample ply.

Then I tried it in plane-opt-rgbd. I got some incomplete textures in the output.

The rgbd frames seemed good without obvious noise. I wonder why this happened for the whole door is close to a plane.

Although there are just a little incomplete textures, the problem can be nonnegligible in my own data.

Hope you can offer some advices about fullfilling the textures. Thank you so much.

Hello, I have used Scannet data(based on Bundelfusion) to run this project, but the texture image includes only one patch.

Hi, I tried the data from the BundleFusion's official site and it worked very well. And I built BundleFusion code on my own. The data generated by my BundleFusion code showed bad texture result in plane-opt-rgbd. I feel puzzled.

The mesh partition and my original mesh seem good. But I dont know where is the problem.

Hope to get your reply, thanks.