Deforming Autoencoders: Unsupervised Disentangling of Shape and Appearance
This work addresses the challenge of unsupervised image analysis for computer vision researchers, offering a novel method for disentangling shape and appearance without labeled data, though it appears incremental as it builds on the deformable template paradigm.
The authors tackled the problem of unsupervised disentangling of shape and appearance in images by introducing Deforming Autoencoders, a generative model that represents shape as a deformation from a template and appearance in canonical coordinates, enabling tasks like expression morphing, face manipulation, and unsupervised landmark localization.
In this work we introduce Deforming Autoencoders, a generative model for images that disentangles shape from appearance in an unsupervised manner. As in the deformable template paradigm, shape is represented as a deformation between a canonical coordinate system (`template') and an observed image, while appearance is modeled in `canonical', template, coordinates, thus discarding variability due to deformations. We introduce novel techniques that allow this approach to be deployed in the setting of autoencoders and show that this method can be used for unsupervised group-wise image alignment. We show experiments with expression morphing in humans, hands, and digits, face manipulation, such as shape and appearance interpolation, as well as unsupervised landmark localization. A more powerful form of unsupervised disentangling becomes possible in template coordinates, allowing us to successfully decompose face images into shading and albedo, and further manipulate face images.