search

ardaduz / deep-video-mvs

Code for "DeepVideoMVS: Multi-View Stereo on Video with Recurrent Spatio-Temporal Fusion" (CVPR 2021)
MIT License
221 stars 29 forks source link

TUM-dataset #5

Closed LifeBeyondExpectations closed 3 years ago

LifeBeyondExpectations commented 3 years ago

Thank you for sharing this wonderful work.

I have a question about the usage of the TUM-dataset. In the paper, it describes that it selectively uses 13 scenes from the TUM dataset. But in the code, it parses 15 scenes. 

input_directories = [
        input_folder / "rgbd_dataset_freiburg1_desk",
        input_folder / "rgbd_dataset_freiburg1_plant",
        input_folder / "rgbd_dataset_freiburg1_room",
        input_folder / "rgbd_dataset_freiburg1_teddy",
        input_folder / "rgbd_dataset_freiburg1_xyz",
        input_folder / "rgbd_dataset_freiburg2_desk",
        input_folder / "rgbd_dataset_freiburg2_metallic_sphere2",
        input_folder / "rgbd_dataset_freiburg2_xyz",
        input_folder / "rgbd_dataset_freiburg3_cabinet",
        input_folder / "rgbd_dataset_freiburg3_long_office_household",
        input_folder / "rgbd_dataset_freiburg3_nostructure_notexture_far",
        input_folder / "rgbd_dataset_freiburg3_nostructure_texture_far",
        input_folder / "rgbd_dataset_freiburg3_structure_notexture_far",
        input_folder / "rgbd_dataset_freiburg3_structure_texture_far",
        input_folder / "rgbd_dataset_freiburg3_teddy"]

I am confused by this inconsistency.

ardaduz commented 3 years ago

Thank you for pointing this mistake out. There is a missing update before sharing the code. The names of the 13 selected scenes are correct in the supplementary material of the paper. I also updated this piece of code in the repo.