on the line 316 of main.py

[wangty537]

thanks the code, on the line 316 of main.py, depth_sil_rendervar = transformed_params2depthplussilhouette(params, curr_data['w2c'], transformed_pts) , tranformed_pts have been already is the current camera corr, but in the function you have transform again as the line213 of EndoGSLAM/utils /slam_helpers.py, could you explain

[Loping151]

It depends on the input data. In our case, transformed_pts are already in the current camera frame, but the passed w2c is the first-frame w2c. Since relative_pose=True, it is effectively an identity matrix, so no actual double transformation occurs. The operation is just redundant.