Thank you for such a great work and code.
However, I found that the pose given by the 7scenes dataset is somewhat different from the pose in 7scenes_sfm_triangulated. For example, for seq5/frame-000000.png, the rotation in the 7scenes dataset is (I use transpose to convert the rotation of c2w to w2c)：
and in 7scenes_sfm_triangulated is:
Are the gt poses of 7scenes_sfm_triangulated more accurate?
Thank you for such a great work and code. However, I found that the pose given by the 7scenes dataset is somewhat different from the pose in 7scenes_sfm_triangulated. For example, for seq5/frame-000000.png, the rotation in the 7scenes dataset is (I use transpose to convert the rotation of c2w to w2c)：
and in 7scenes_sfm_triangulated is:
Are the gt poses of 7scenes_sfm_triangulated more accurate?