Generative Visual Manipulation on the Natural Image Manifold

In ECCV'16

People

• Jun-Yan Zhu
• Philipp Krähenbühl
• Eli Shechtman (Adobe)
• Alexei A. Efros

Popular Press

Abstract

Realistic image manipulation is challenging because it requires modifying the image appearance in a user-controlled way, while preserving the realism of the result. Unless the user has considerable artistic skill, it is easy to "fall off" the manifold of natural images while editing. In this paper, we propose to learn the natural image manifold directly from data using a generative adversarial neural network. We then define a class of image editing operations, and constrain their output to lie on that learned manifold at all times. The model automatically adjusts the output keeping all edits as realistic as possible. All our manipulations are expressed in terms of constrained optimization and are applied in near-real time. We evaluate our algorithm on the task of realistic photo manipulation of shape and color. The presented method can further be used for changing one image to look like the other, as well as generating novel imagery from scratch based on user's scribbles.

Paper

ECCV 2016 paper, 6.2MB

Citation

Jun-Yan Zhu, Philipp Krähenbühl, Eli Shechtman and Alexei A. Efros. "Generative Visual Manipulation on the Natural Image Manifold", in European Conference on Computer Vision (ECCV). 2016. Bibtex

Video

Video (Download), 74MB

Code and Data: Github

Interactive Image Generation Our system can create new imagery based on user's scribbles. Let's draw an outdoor scene.

Here we demonstrate how to achieve the above result.

We can use our drawing interface to design products as well.

 

Intelligent Image Editing Our interactive system allows a user to manipulate image in a natural and realistic way.

 

Generative Image Transformation Our system can automatically transform the shape and color of one image to look like another image.

Related Work

• Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, Yoshua Bengio "Generative Adversarial Networks", in NIPS 2014.
• Alec Radford, Luke Metz and Soumith Chintala "Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks", in ICLR 2016.
• Jun-Yan Zhu, Yong Jae Lee and Alexei A. Efros. "AverageExplorer: Interactive Exploration and Alignment of Visual Data Collections", in SIGGRAPH 2014.
• Jun-Yan Zhu, Philipp Krähenbühl, Eli Shechtman and Alexei A. Efros. "Learning a Discriminative Model for the Perception of Realism in Composite Images", in ICCV 2015.

Funding

This research is supported in part by:

• Adobe research grant.
• eBay research grant.
• Intel research grant.
• NVIDIA hardware donation.
• Jun-Yan Zhu is supported by Facebook Graduate Fellowship.