Semantic Image Manipulation Using Scene Graphs (CVPR 2020)

Helisa Dhamo *     Azade Farshad *     Iro Laina     Nassir Navab     Gregory D. Hager     Federico Tombari     Christian Rupprecht    

Technical University of Munich    University of Oxford    Johns Hopkins University    Google

* The first two authors contributed equally.

In our work, we address the novel problem of image manipulation from scene graphs, in which a user can edit images by merely applying changes in the nodes or edges of a semantic graph that is generated from the image. Our goal is to encode image information in a given constellation and from there on generate new constellations, such as replacing objects or even changing relationships between objects, while respecting the semantics and style from the original image. We introduce a spatio-semantic scene graph network that does not require direct supervision for constellation changes or image edits. This makes it possible to train the system from existing real-world datasets with no additional annotation effort.


Conference on Computer Vision and Pattern Recognition (CVPR) 2020
Paper PDF | arXiv
    title={Semantic Image Manipulation Using Scene Graphs},
    author={Dhamo, Helisa and Farshad, Azade, and Laina, Iro and Navab, Nassir and
            Hager, Gregory D., and Tombari, Federico and Rupprecht, Christian},


We provide the learned Pytorch checkpoints for Visual Genome and CLEVR.

The predicted VG scene graphs from
Factorizable Net, used in our experiments can be downloaded here.

The automatically generated dataset with editing pairs, based on CLEVR, can be downloaded

Source Code

The source code for this work can be found here.


For questions regarding the method, code, or the CLEVR data, please contact us.