Skip to content

README.md

Latest commit

 

History

History
46 lines (31 loc) · 2.63 KB

README.md

File metadata and controls

46 lines (31 loc) · 2.63 KB

Generate dataset for the Advanced environments

  1. Download the 3D Front and 3D Future dataset upon request here.

  2. (Optional) Simplify the OBJ meshes in the 3D Future dataset (16565 pieces of furniture) using Blender script from Bakri - this can take a while:

    sudo apt-get install blender
python data/process/advanced/simplify_mesh.py --mesh_folder=[path_to_3D_FUTURE_MODEL] --num_cpus=.

  3. (Optional) Downsize the texture image files in the 3D-Front dataset:

    python data/process/advanced/downsize_texture.py --texture_folder=[path_to_3D_FRONT_TEXTURE] --img_size=128 --num_cpus=.

  4. (For the Advanced-Dense setting), generate >2000 room meshes by randomly sampling furniture meshes from the 3D Future dataset, and save the task configurations in a new folder - this can take a while:

    python data/process/advanced/generate_advanced_dense.py --num_cpus=. --save_task_folder=[path_to_dense_task] --mesh_folder=[path_to_3D_FUTURE_MODEL] --texture_folder=[path_to_3D_FRONT_TEXTURE] --floor_mtl_path=. --wall_mtl_path=. --use_simplified_mesh=. --num_room=2500

  5. (For the Advanced-Realistic setting), generate 6813 house meshes by reading the json files in the 3D Front dataset and merging the meshes from the 3D Future dataset, and save them in a new folder:

    python data/process/advanced/json_to_obj.py --num_cpus=. --save_folder=[house_folder] --future_folder=[path_to_3D_FUTURE_MODEL] --json_folder=[path_to_3D_FRONT] --texture_folder=[path_to_3D_FRONT_TEXTURE] --floor_mtl_path=. --wall_mtl_path=. --use_simplified_mesh=.

  6. (For the Advanced-Realistic setting), post-process the house meshes, and save the task configurations in a new folder - this can take a while and requires a siginicant amount of RAM. Roughly 2500 configurations would be saves.

    python data/process/advanced/generate_advanced_realistic.py --num_cpus=. --save_task_folder=[path_to_realistic_task] --house_folder=[house_folder]

  7. (For the Advanced-Dense setting), split the generated task configurations to train and test datasets.

    python data/process/advanced/split_advanced_dense.py --save_folder=[path_to_dense_dataset] --task_folder=[path_to_dense_task]

  8. (For the Advanced-Realistic setting), split the generated task configurations to train and test datasets.

    python data/process/advanced/split_advanced_dense.py --save_folder=[path_to_realistic_dataset] --task_folder=[path_to_realistic_task]