The position of reconstructed mesh in quali_eval

[lw-ethan]

Hi, thanks for your nice work and concise repo!

I'm working on the project that needs to leverage on the reconstructed mesh generated from dbw, and I found the position and scale of mesh have certain gap compared to points computed from colmap.
I'm wondering if the mesh is not placed on the real position of world coordinates, and instead always placed at the origin?

[monniert]

Hi @lw-ethan, the scene is indeed always placed at the origin. It is transformed to the colmap space using a predefined rotation, translation and scale matrix

[lw-ethan]

@monniert, thanks for your reply, I found that it was my mistake:
I kept stuck at the configuration of S_world, R_world, T_world, which should be obtained easily by delta transformation in Blender
I came to realize later the problem arise from the settings of R_world, the mesh applies world coordinate this way verts = (verts * self.S_world) @ self.R_world + self.T_world[:, None], the vertices are at the left side of R_world
However I used the convention that R = RzRyRx * verts, which caused the transformation wasn't as my expect
Changing the conversion R_world = (elev_to_rotation_matrix(elev) @ azim_to_rotation_matrix(azim) @ roll_to_rotation_matrix(roll))[None] to R_world = (roll_to_rotation_matrix(roll) @ azim_to_rotation_matrix(azim) @ elev_to_rotation_matrix(elev)).T[None] met my assumption and solved my problem