Context Encoders: Feature Learning by Inpainting

TL;DR

This paper introduces Context Encoders, a convolutional neural network trained to predict missing parts of images based on surrounding context, enabling unsupervised feature learning and applications in inpainting and downstream vision tasks.

Contribution

The paper proposes a novel unsupervised learning method using context encoders with adversarial training to produce sharp, semantically meaningful image inpainting results.

Findings

Learned features improve CNN pre-training for classification, detection, and segmentation.

Adversarial loss yields sharper and more plausible inpainting results.

Context encoders capture both appearance and semantic information of visual structures.

Abstract

We present an unsupervised visual feature learning algorithm driven by context-based pixel prediction. By analogy with auto-encoders, we propose Context Encoders -- a convolutional neural network trained to generate the contents of an arbitrary image region conditioned on its surroundings. In order to succeed at this task, context encoders need to both understand the content of the entire image, as well as produce a plausible hypothesis for the missing part(s). When training context encoders, we have experimented with both a standard pixel-wise reconstruction loss, as well as a reconstruction plus an adversarial loss. The latter produces much sharper results because it can better handle multiple modes in the output. We found that a context encoder learns a representation that captures not just appearance but also the semantics of visual structures. We quantitatively demonstrate the effectiveness of our learned features for CNN pre-training on classification, detection, and segmentation tasks. Furthermore, context encoders can be used for semantic inpainting tasks, either stand-alone or as initialization for non-parametric methods.