Home

Awesome

ID-Unet: Iterative-view-synthesis(CVPR2021 Oral)

Tensorflow implementation of ID-Unet: Iterative Soft and Hard Deformation for View Synthesis.

Overview architecture

<p align="center"> <img src="./results/overview.png" width="55%"><br><center></center></p>

The network architecture

<p align="center"> <img src="./results/architecture.png" width="95%"><br><center></center></p> <p align="center"> <img src="./results/CDM.png" width="95%"><br><center></center></p>

Experiment Results

<p align="center"> <img src="./results/result1.png" width="55%"><br><center></center></p> <p align="center"> <img src="./results/result3.png" width="95%"><br><center></center></p> <p align="center"> <img src="./results/result2.png" width="95%"><br><center></center></p> <p align="center"> <img src="./results/result4.png" width="65%"><br><center></center></p> <p align="center"> <img src="./results/flow.png" width="65%"><br><center></center></p>

Preparation

Quick Start

Exemplar commands are listed here for a quick start.

dataset

Training

Testing

Citation

If this work is useful for your research, please consider citing:

@inproceedings{yin2021id, title={ID-Unet: Iterative Soft and Hard Deformation for View Synthesis}, author={Yin, Mingyu and Sun, Li and Li, Qingli}, booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition}, pages={7220--7229}, year={2021} }