Learning to Manipulate Individual Objects in an Image

Abstract

We describe a method to train a generative model with latent factors that are (approximately) independent and localized. This means that perturbing the latent variables affects only local regions of the synthesized image, corresponding to objects. Unlike other unsupervised generative models, ours enables object-centric manipulation, without requiring object-level annotations, or any form of annotation for that matter. The key to our method is the combination of spatial disentanglement, enforced by a Contextual Information Separation loss, and perceptual cycle-consistency, enforced by a loss that penalizes changes in the image partition in response to perturbations of the latent factors. We test our method's ability to allow independent control of spatial and semantic factors of variability on existing datasets and also introduce two new ones that highlight the limitations of current methods.

References in corpus (9)

• Auto-Encoding Variational Bayes
• Deep Convolutional Inverse Graphics Network
• Understanding disentangling in $\beta$-VAE
• MONet: Unsupervised Scene Decomposition and Representation
• Multi-Level Variational Autoencoder: Learning Disentangled Representations from Grouped Observations
• Sequential Attend, Infer, Repeat: Generative Modelling of Moving Objects
• 3D-Aware Scene Manipulation via Inverse Graphics
• Object Discovery with a Copy-Pasting GAN
• Emerging Disentanglement in Auto-Encoder Based Unsupervised Image Content Transfer

Cited by in corpus (1)

• Object-Centric Learning with Slot Attention