# DiffuScene [**Paper**](https://arxiv.org/abs/2303.14207.pdf) | [**arXiv**](https://arxiv.org/abs/2303.14207.pdf) | [**Video**](https://www.youtube.com/embed/PFi-03T1lN0) | [**Project Page**](https://tangjiapeng.github.io/projects/DiffuScene/)
This is the repository that contains source code for the paper: **DiffuScene: Denoising Diffusion Models for Generative Indoor Scene Synthesis** - We present DiffuScene, a diffusion model for diverse and realistic indoor scene synthesis.
## Installation & Dependencies You can create a conda environment called ```diffuscene``` using ``` conda env create -f environment.yaml conda activate diffuscene ``` Next compile the extension modules. You can do this via ``` python setup.py build_ext --inplace pip install -e . ``` Install ChamferDistancePytorch ``` cd ChamferDistancePytorch/chamfer3D python setup.py install ``` ## Download The pretrained model, results, and preprocess datasets are put into [GoogleDrive](https://drive.google.com/drive/folders/1EhvyNCAWWto6vMt0vXWMKBoSdYR_9pC2?usp=sharing) ### Pretrained models and Results The pretrained models of DiffuScene (unconditional, text, scene_rearrangement) can be downloaded from [here](https://drive.google.com/file/d/1pk9AzGcBz_kRfmRzvFNDW5byk4MwbXEm/view?usp=sharing). The pretrained model of ShapeAutoencoder can be downloaded from [here](https://drive.google.com/file/d/1xuFa_Hh6BOZqfaAnlG-W_9mA1gilU6nI/view?usp=sharing). The google drive also contains [the results of unconditional generation](https://drive.google.com/file/d/1EJ2VPZvbuKeR-ob1SsiwEC7cAEFDug3i/view?usp=sharing). #### Preprocessed dataset To directly train and evaluate DiffuScene without tedious data preprocessing, we also provide the preprocessed [3D-Front](https://drive.google.com/file/d/1UNSFN0kULyOzUErDPVvkKYbmfzA-4MsG/view?usp=sharing) and [3D-FUTURE](https://drive.google.com/file/d/16fz81Eh6B6pbZNGTMbe0FdfNwIuQicE1/view?usp=sharing) ## Dataset The training and evaluation are based on the [3D-FRONT](https://tianchi.aliyun.com/specials/promotion/alibaba-3d-scene-dataset) and the [3D-FUTURE](https://www.google.com/search?q=3d-future&oq=3d-fut&aqs=chrome.1.69i57j0j0i30l8.3909j0j7&sourceid=chrome&ie=UTF-8) dataset. To download both datasets, please refer to the instructions provided in the dataset's [webpage](https://tianchi.aliyun.com/specials/promotion/alibaba-3d-scene-dataset). ### Pickle the 3D-FUTURE dataset To accelerate the preprocessing speed, we can sepcify the `PATH_TO_SCENES` environment variable for all scripts. This filepath contains the parsed `ThreedFutureDataset` after being pickled. To pickle it, you can simply run this script as follows: ``` python pickle_threed_future_dataset.py path_to_output_dir path_to_3d_front_dataset_dir path_to_3d_future_dataset_dir path_to_3d_future_model_info --dataset_filtering room_type ``` Based on the pickled ThreedFutureDataset, we also provide a script to pickle the sampled point clouds of object CAD models, which are used to shape autoencoder training and latent shape code extraction. ``` python pickle_threed_future_pointcloud.py path_to_output_dir path_to_3d_front_dataset_dir path_to_3d_future_dataset_dir path_to_3d_future_model_info --dataset_filtering room_type ``` For example, ``` python pickle_threed_future_dataset.py /cluster/balrog/jtang/3d_front_processed/ /cluster/balrog/jtang/3D-FRONT/ /cluster/balrog/jtang/3D-FUTURE-model /cluster/balrog/jtang/3D-FUTURE-model/model_info.json --dataset_filtering threed_front_livingroom --annotation_file ../config/livingroom_threed_front_splits.csv PATH_TO_SCENES="/cluster/balrog/jtang/3d_front_processed/threed_front.pkl" python pickle_threed_fucture_pointcloud.py /cluster/balrog/jtang/3d_front_processed/ /cluster/balrog/jtang/3D-FRONT/ /cluster/balrog/jtang/3D-FUTURE-model /cluster/balrog/jtang/3D-FUTURE-model/model_info.json --dataset_filtering threed_front_livingroom --annotation_file ../config/livingroom_threed_front_splits.csv ``` Note that these two scripts should be separately executed for different room types containing different objects. For the case of 3D-FRONT this is for the bedrooms and the living/dining rooms, thus you have to run this script twice with different `--dataset_filtering` and `--annotation_file`options. Please check the help menu for additional details. ### Train shape autoencoder Then you can train the shape autoencoder using all models from bedrooms/diningrooms/livingrooms. ``` cd ./scripts PATH_TO_SCENES="/cluster/balrog/jtang/3d_front_processed/threed_front.pkl" python train_objautoencoder.py ../config/obj_autoencoder/bed_living_diningrooms_lat32.yaml your_objae_output_directory --experiment_tag "bed_living_diningrooms_lat32" --with_wandb_logger ``` ### Pickle Latent Shape Code Next, you can use the pre-train checkpoint of shape autoencoder to extract latent shape codes for each room type. Take the bedrooms for example: ``` PATH_TO_SCENES="/cluster/balrog/jtang/3d_front_processed/threed_front.pkl" python generate_objautoencoder.py ../config/objautoencoder/bedrooms.yaml your_objae_output_directory --experiment_tag "bed_living_diningrooms_lat32" ``` ### Preprocess 3D-Front dataset with latent shape codes Finally, you can run `preprocessing_data.py` to read and pickle object properties (class label, location, orientation, size, and latent shape features) of each scene. ``` PATH_TO_SCENES="/cluster/balrog/jtang/3d_front_processed/threed_front.pkl" python preprocess_data.py /cluster/balrog/jtang/3d_front_processed/livingrooms_objfeats_32_64 /cluster/balrog/jtang/3D-FRONT/ /cluster/balrog/jtang/3D-FUTURE-model /cluster/balrog/jtang/3D-FUTURE-model/model_info.json --dataset_filtering threed_front_livingroom --annotation_file ../config/livingroom_threed_front_splits.csv --add_objfeats ``` ## Training & Evaluate Diffuscene To train diffuscene on 3D Front-bedrooms, you can run ``` ./run/train.sh ./run/train_text.sh ``` To generate the scene of unconditional and text-conditioned scene generation with our pretraiened models, you can run ``` ./run/generate.sh ./run/generate_text.sh ``` If you want to calculate evaluation metrics of bbox IoU and average number of symmetric pairs, you can add the option```--compute_intersec```. Please note that our current text-conditioned model is used to generate a full scene configuration from a text prompt of partial scene (2-3 sentences). If you want to evaluate our method with text prompts of more sentences, you might need to re-train our method. ## Evaluation Metrics To evaluate FID and KID from rendered 2D images of generated and reference scenes, you can run: ``` python compute_fid_scores.py $ground_truth_bedrooms_top2down_render_folder $generate_bedrooms_top2down_render_folder ../config/bedroom_threed_front_splits.csv python compute_fid_scores.py $ground_truth_diningrooms_top2down_render_folder $generate_diningrooms_top2down_render_folder ../config/diningroom_threed_front_splits.csv ``` To evaluate improved precision and recall, you can run: ``` python improved_precision_recall.py $ground_truth_bedrooms_top2down_render_folder $generate_bedrooms_top2down_render_folder ../config/bedroom_threed_front_splits.csv python improved_precision_recall.py $ground_truth_diningrooms_top2down_render_folder $generate_diningrooms_top2down_render_folder ../config/diningroom_threed_front_splits.csv ``` ## Relevant Research Please also check out the following papers that explore similar ideas: - LEGO-Net: Learning Regular Rearrangements of Objects in Rooms.[[homepage]](https://ivl.cs.brown.edu/research/lego-net.html) - Learning 3D Scene Priors with 2D Supervision. [[homepage]](https://yinyunie.github.io/sceneprior-page/) - Sceneformer: Indoor Scene Generation with Transformers. [[homepage]](https://xinpeng-wang.github.io/sceneformer/) - ATISS: Autoregressive Transformers for Indoor Scene Synthesis. [[homepage]](https://research.nvidia.com/labs/toronto-ai/ATISS/) - Scene Synthesis via Uncertainty-Driven Attribute Synchronization [[pdf]](https://openaccess.thecvf.com/content/ICCV2021/papers/Yang_Scene_Synthesis_via_Uncertainty-Driven_Attribute_Synchronization_ICCV_2021_paper.pdf) - Indoor Scene Generation from a Collection of Semantic-Segmented Depth Images [[pdf]](https://openaccess.thecvf.com/content/ICCV2021/papers/Yang_Indoor_Scene_Generation_From_a_Collection_of_Semantic-Segmented_Depth_Images_ICCV_2021_paper.pdf) - Fast and Flexible Indoor Scene Synthesis via Deep Convolutional Generative Models [[pdf]](https://arxiv.org/pdf/1811.12463.pdf) ## Citation If you find DiffuScene useful for your work please cite: ``` @inproceedings{tang2024diffuscene, title={Diffuscene: Denoising diffusion models for generative indoor scene synthesis}, author={Tang, Jiapeng and Nie, Yinyu and Markhasin, Lev and Dai, Angela and Thies, Justus and Nie{\ss}ner, Matthias}, booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition}, year={2024} } ``` Contact [Jiapeng Tang](mailto:tangjiapengtjp@gmail.com) for questions, comments and reporting bugs. ## Acknowledgements Most of the code is borrowed from [ATISS](https://github.com/nv-tlabs/ATISS). We thank for Despoina Paschalidou her great works and repos.