Overview

This is the code repository for LaLaLoc and LaLaLoc++.

• We currently provide:
  • Training and evaluation code for LaLaLoc, for both the Image-to-Layout and Layout-to-Layout configurations.
  • Training and evaluation code for LaLaLoc++'s plan and image branches.
  • Pretrained models for all the provided configs.

LaLaLoc++: Global Floor Plan Comprehension for Layout Localisation in Unvisited Environments

Henry Howard-Jenkins and Victor Adrian Prisacariu (ECCV 2022)

Project Page | Paper(coming soon!)

LaLaLoc: Latent Layout Localisation in Dynamic, Unvisited Environments

Henry Howard-Jenkins, Jose-Raul Ruiz-Sarmiento and Victor Adrian Prisacariu (ICCV 2021)

Project Page | Paper

Setup

Installing Requirements

• Create conda environment:

conda create -n lalaloc python==3.8
conda activate lalaloc

• Install PyTorch:

conda install pytorch==1.7.1 torchvision==0.8.2 cudatoolkit=10.1 -c pytorch

• Install Pytorch Lightning:

conda install -c conda-forge pytorch-lightning==1.1.5

• Install Pytorch3d:

conda install -c fvcore -c iopath -c conda-forge fvcore iopath
conda install -c bottler nvidiacub
conda install -c pytorch3d pytorch3d==0.4.0 

• Install Pymesh
  • Follow build and install instructions: https://github.com/PyMesh/PyMesh
• Install Redner and OpenCV:

pip install redner-gpu opencv-python

• Install Scikit-Learn:

conda install -c anaconda scikit-learn

Download the Structured3D Dataset

• Information provided here: https://github.com/bertjiazheng/Structured3D

Usage

Layout/Plan Branch

• Train LaLaLoc's layout branch or LaLaLoc++'s plan branch.

# LaLaLoc layout branch
python train.py -c configs/layout_branch.yaml \
    DATASET.PATH [path/to/dataset]

# LaLaLoc++ plan branch
python train.py -c configs/lalaloc_pp/plan_branch.yaml \
    DATASET.PATH [path/to/dataset]

• Test LaLaLoc's layout branch:

  • Perform evaluation of the trained layout branch on a sampled grid of 0.5m with VDR and LPO.

  Note: Testing LaLaLoc++'s plan branch isn't particularly meaningful.

python train.py -c configs/layout_branch.yaml -t [path/to/checkpoint] \
    DATASET.PATH [path/to/dataset] \
    SYSTEM.NUM_GPUS 1 \
    TEST.VOGEL_DISC_REFINE True \
    TEST.LATENT_POSE_OPTIMISATION True \
    TEST.POSE_SAMPLE_STEP 500

Image Branch

• Train the image branch for LaLaLoc and LaLaLoc++
  • Perform training of the image branch with the layout/plan branch from a previous training run.

# LaLaLoc image branch
python train.py -c configs/image_branch.yaml \
    DATASET.PATH [path/to/dataset] \
    TRAIN.SOURCE_WEIGHTS [path/to/layout_branch_checkpoint]

# LaLaLoc++ image branch
python train.py -c configs/lalaloc_pp/image_branch.yaml \
    DATASET.PATH [path/to/dataset] \
    TRAIN.SOURCE_WEIGHTS [path/to/plan_branch_checkpoint]

• Test image branch

# LaLaLoc image branch
python train.py -c configs/image_branch.yaml -t [path/to/checkpoint] \
    DATASET.PATH [path/to/dataset] \
    SYSTEM.NUM_GPUS 1 \
    TEST.VOGEL_DISC_REFINE True \
    TEST.LATENT_POSE_OPTIMISATION True \
    TEST.POSE_SAMPLE_STEP 500

# LaLaLoc++ image branch
python train.py -c configs/lalaloc_pp/transfomer_image_branch.yaml -t [path/to/checkpoint] \
    DATASET.PATH [path/to/dataset] \
    SYSTEM.NUM_GPUS 1 \

Citations

@article{howard2022lalaloc++,
  title={LaLaLoc++: Global Floor Plan Comprehension for Layout Localisation in Unvisited Environments},
  author={Howard-Jenkins, Henry and Prisacariu, Victor Adrian},
  booktitle={Proceedings of the European Conference on Computer Vision},
  pages={},
  year={2022}
}

@inproceedings{howard2021lalaloc,
  title={Lalaloc: Latent layout localisation in dynamic, unvisited environments},
  author={Howard-Jenkins, Henry and Ruiz-Sarmiento, Jose-Raul and Prisacariu, Victor Adrian},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={10107--10116},
  year={2021}
}