GENESIS-V2: Inferring Unordered Object Representations without Iterative Refinement

Abstract

Advances in unsupervised learning of object-representations have culminated in the development of a broad range of methods for unsupervised object segmentation and interpretable object-centric scene generation. These methods, however, are limited to simulated and real-world datasets with limited visual complexity. Moreover, object representations are often inferred using RNNs which do not scale well to large images or iterative refinement which avoids imposing an unnatural ordering on objects in an image but requires the a priori initialisation of a fixed number of object representations. In contrast to established paradigms, this work proposes an embedding-based approach in which embeddings of pixels are clustered in a differentiable fashion using a stochastic stick-breaking process. Similar to iterative refinement, this clustering procedure also leads to randomly ordered object representations, but without the need of initialising a fixed number of clusters a priori. This is used to develop a new model, GENESIS-v2, which can infer a variable number of object representations without using RNNs or iterative refinement. We show that GENESIS-v2 performs strongly in comparison to recent baselines in terms of unsupervised image segmentation and object-centric scene generation on established synthetic datasets as well as more complex real-world datasets.

NeurIPS 2021 camera-ready version; 26 pages, 19 figures

References in corpus (12)

• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Semantic Instance Segmentation with a Discriminative Loss Function
• Semantic Instance Segmentation via Deep Metric Learning
• MONet: Unsupervised Scene Decomposition and Representation
• On the Binding Problem in Artificial Neural Networks
• Learning Physical Graph Representations from Visual Scenes
• Object Discovery with a Copy-Pasting GAN
• RELATE: Physically Plausible Multi-Object Scene Synthesis Using Structured Latent Spaces
• SCALOR: Generative World Models with Scalable Object Representations
• Greedy Hierarchical Variational Autoencoders for Large-Scale Video Prediction
• Reconstruction Bottlenecks in Object-Centric Generative Models
• Multi-objects Generation with Amortized Structural Regularization

Cited by in corpus (5)

• Object-aware Contrastive Learning for Debiased Scene Representation
• Generalization and Robustness Implications in Object-Centric Learning
• Unsupervised Object Learning via Common Fate
• Leveraging Hidden Structure in Self-Supervised Learning
• Illiterate DALL-E Learns to Compose