fivecoco / floornet

FloorNet: A Unified Framework for Floorplan Reconstruction from 3D Scans

By Chen Liu^*, Jiaye Wu^*, and Yasutaka Furukawa (^* indicates equal contribution)

Introduction

This paper proposes FloorNet, a novel neural network, to turn RGBD videos of indoor spaces into vector-graphics floorplans. FloorNet consists of three branches, PointNet branch, Floorplan branch, and Image branch. For more details, please refer to our Arxiv paper or visit our project website. This is a follow-up work of our floorplan transformation project which you can find here.

Dependencies

Python 2.7, TensorFlow (>= 1.0), numpy, opencv 3.

Data

We collect 155 scans of residential units and annotated corresponding floorplan information. Among 155 scans, 135 are used for training and 20 are for testing. We convert both training data and testing data to tfrecords files which can be downloaded here. Please put the downloaded files under folder data/.

Training

To train the network from scratch, please run:

python train.py --restore=0

Evaluation

To evaluate the performance of our trained model, please run:

python train.py --task=evaluate

Contact

If you have any questions, please contact me at [email protected].