Counterfactual Image Networks


Nov 03, 2017 (modified: Nov 03, 2017) ICLR 2018 Conference Blind Submission readers: everyone Show Bibtex
  • Abstract: We capitalize on the natural compositional structure of images in order to learn object segmentation with unlabeled images. The intuition behind our approach is that removing objects from images will yield natural images, however removing random patches will yield unnatural images. We leverage this signal to develop a generative model that decomposes an image into layers, and when all layers are combined, it reconstructs the input image. However, when a layer is removed, the model learns to produce a different image that still looks natural to an adversary, which is possible by removing objects. Experiments and visualizations suggest that this model automatically learns object segmentation better than baselines.
  • TL;DR: Unsupervised image segmentation using compositional structure of images and generative models.
  • Keywords: computer vision, image segmentation, generative models, adversarial networks, unsupervised learning