Learning Physical Graph Representations from Visual Scenes

by   Daniel M. Bear, et al.

Convolutional Neural Networks (CNNs) have proved exceptional at learning representations for visual object categorization. However, CNNs do not explicitly encode objects, parts, and their physical properties, which has limited CNNs' success on tasks that require structured understanding of visual scenes. To overcome these limitations, we introduce the idea of Physical Scene Graphs (PSGs), which represent scenes as hierarchical graphs, with nodes in the hierarchy corresponding intuitively to object parts at different scales, and edges to physical connections between parts. Bound to each node is a vector of latent attributes that intuitively represent object properties such as surface shape and texture. We also describe PSGNet, a network architecture that learns to extract PSGs by reconstructing scenes through a PSG-structured bottleneck. PSGNet augments standard CNNs by including: recurrent feedback connections to combine low and high-level image information; graph pooling and vectorization operations that convert spatially-uniform feature maps into object-centric graph structures; and perceptual grouping principles to encourage the identification of meaningful scene elements. We show that PSGNet outperforms alternative self-supervised scene representation algorithms at scene segmentation tasks, especially on complex real-world images, and generalizes well to unseen object types and scene arrangements. PSGNet is also able learn from physical motion, enhancing scene estimates even for static images. We present a series of ablation studies illustrating the importance of each component of the PSGNet architecture, analyses showing that learned latent attributes capture intuitive scene properties, and illustrate the use of PSGs for compositional scene inference.


page 2

page 7

page 8

page 9

page 17

page 20

page 21


Learning and generalization of compositional representations of visual scenes

Complex visual scenes that are composed of multiple objects, each with a...

Generative Hierarchical Models for Parts, Objects, and Scenes

Compositional structures between parts and objects are inherent in natur...

Prediction of Scene Plausibility

Understanding the 3D world from 2D images involves more than detection a...

Object-Level Context Modeling For Scene Classification with Context-CNN

Convolutional Neural Networks (CNNs) have been used extensively for comp...

Spatial Mixture Models with Learnable Deep Priors for Perceptual Grouping

Humans perceive the seemingly chaotic world in a structured and composit...

Physion: Evaluating Physical Prediction from Vision in Humans and Machines

While machine learning algorithms excel at many challenging visual tasks...

Searching Scenes by Abstracting Things

In this paper we propose to represent a scene as an abstraction of 'thin...

Please sign up or login with your details

Forgot password? Click here to reset