chengjieniu / Im2Struct

Im2Struct: Recovering 3D Shape Structure from a Single RGB Image

By Chengjie Niu, Jun Li, Kai Xu.

This repository contains the pre-trained models and testing code for recovering 3D shape structures from single RGB images.

Citation

If you find our work useful in your research, please consider citing:

@inProceedings{niu_cvpr18,
title={Im2Struct: Recovering 3D Shape Structure from a Single RGB Image},
author = {Chengjie Niu
and Jun Li
and Kai Xu},
booktitle={Computer Vision and Pattern Regognition (CVPR)},
year={2018}
}

Guide:

Our current release has been tested on Matlab 2015b

1. Download and compile MatConvNet(http://www.vlfeat.org/matconvnet/)

run matlab/vl_setupnn;   

vl_compilenn('enableGpu', true, ... 
	     'cudaRoot', '/Developer/NVIDIA/CUDA-X.X', ... 
	     'cudaMethod', 'nvcc');

2. Download our pre-trained model

Please download the pre-trained model according to the description in model/download.txt.

3. Run im2struct_demo.m

Run im2struct_demo.m to recover 3D shape structure from an single RGB image with our pre-trained model. The recovered 3D shape structures can be visulized in Matlab.

The recovered 3D shape structures for example images(data/example_1/2.jpg) should look as follows:

For any questions, please contact Chengjie Niu([e-mail]).