This repository is the offical tensorflow implementation of MGTnet:Image-Guided Human Reconstruction via Multi-Scale Graph Transformation Networks(TIP 2020).

3D human reconstruction from a single image is a challenging problem. Existing methods have difficulties to infer 3D clothed human models with consistent topologies for various poses. In this paper, we propose an efficient and effective method using a hierarchical graph transformation network. To deal with large deformations and avoid distorted geometries, rather than using Euclidean coordinates directly, 3D human shapes are represented by a vertex-based deformation representation that effectively encodes the deformation and copes well with large deformations. To infer a 3D human mesh consistent with the input real image, we also use a perspective projection layer to incorporate perceptual image features into the deformation representation. Our model is easy to train and fast to converge with short test time.

We present the D^2Human (Dynamic Detailed Human) dataset, including variously posed 3D human meshes with consistent topologies and rich geometry details, together with the captured color images and SMPL. BaiDuYunPan Download with extraction code 69ba.

• Python3.6
• tensorflow1.13.2
• Anaconda

Create virtual environment

conda create -n MGTnet python=3.6
conda activate MGTnet

Install cudn and cudnn

conda install cudatoolkit=10.0 cudnn=7.6.4

Install tensorflow

pip install tensorflow-gpu==1.13.2

Install

Install other environments

pip install -r requirements.txt 

python main_tfrecord.py --mode test

data_preparsion...

python main_tfrecord.py --mode train

Please considering citing

@article{li2021image,
  title={Image-Guided Human Reconstruction via Multi-Scale Graph Transformation Networks},
  author={Li, Kun 
  and Wen, Hao 
  and Feng, Qiao 
  and Zhang, Yuxiang 
  and Li, Xiongzheng 
  and Huang, Jing 
  and Yuan, Cunkuan 
  and Lai, Yu-Kun 
  and Liu, Yebin},
  journal={IEEE Transactions on Image Processing},
  volume={30},
  pages={5239--5251},
  year={2021},
  publisher={IEEE}
}