/BridgeGAN

Code for our 3DV18 paper -- "Generative Adversarial Frontal View to Bird View Synthesis"

Primary LanguagePythonOtherNOASSERTION

Generative Adversarial Front-view to Birdview Synthesis

Top: (Frontal View)
Left: (GroundTruth); Middle: (Ours); Right: (Homography)

Descriptions


More detailed description about the paper can be found in the slides.

Supplementary Materials can be found in this link

TODO List


  • Dataset
  • Code (Tuning)
  • Demo

Dataset


We provide the pre-processed dataset, including the cropped frontal images, homography images and birdview images. You can download it in this URL.

Paper


@inproceedings{zhu2018generative,
  title={Generative Adversarial Frontal View to Bird View Synthesis},
  author={Zhu, Xinge and Yin, Zhichao and Shi, Jianping and Li, Hongsheng and Lin, Dahua},
  booktitle={2018 International Conference on 3D Vision (3DV)},
  pages={454--463},
  year={2018},
  organization={IEEE}
}

Reference


We thanks for the open-source codebase -- CoGAN and UNIT.

License


  • the CC BY-NC-ND 4.0 license