Scene Representation Networks: Continuous 3D-Structure-Aware Neural Scene Representations

Abstract

Unsupervised learning with generative models has the potential of discovering rich representations of 3D scenes. While geometric deep learning has explored 3D-structure-aware representations of scene geometry, these models typically require explicit 3D supervision. Emerging neural scene representations can be trained only with posed 2D images, but existing methods ignore the three-dimensional structure of scenes. We propose Scene Representation Networks (SRNs), a continuous, 3D-structure-aware scene representation that encodes both geometry and appearance. SRNs represent scenes as continuous functions that map world coordinates to a feature representation of local scene properties. By formulating the image formation as a differentiable ray-marching algorithm, SRNs can be trained end-to-end from only 2D images and their camera poses, without access to depth or shape. This formulation naturally generalizes across scenes, learning powerful geometry and appearance priors in the process. We demonstrate the potential of SRNs by evaluating them for novel view synthesis, few-shot reconstruction, joint shape and appearance interpolation, and unsupervised discovery of a non-rigid face model.

Video: https://youtu.be/6vMEBWD8O20 Project page: https://vsitzmann.github.io/srns/

References in corpus (6)

• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Conditional Image Generation with PixelCNN Decoders
• Pixel Recurrent Neural Networks
• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• DRAW: A Recurrent Neural Network For Image Generation
• Transformation Properties of Learned Visual Representations

Cited by in corpus (16)

• DONeRF: Towards Real-Time Rendering of Compact Neural Radiance Fields using Depth Oracle Networks
• Implicit Full Waveform Inversion with Deep Neural Representation
• Joint Implicit Image Function for Guided Depth Super-Resolution
• CLIP-Fields: Weakly Supervised Semantic Fields for Robotic Memory
• From SLAM to Situational Awareness: Challenges and Survey
• Neural Human Performer: Learning Generalizable Radiance Fields for Human Performance Rendering
• Dex-NeRF: Using a Neural Radiance Field to Grasp Transparent Objects
• PIFuHD: Multi-Level Pixel-Aligned Implicit Function for High-Resolution 3D Human Digitization
• Coordinate-based neural representations for computational adaptive optics in widefield microscopy
• Rethinking Positional Encoding
• Equivariant Neural Rendering
• A Survey on Deep Generative 3D-aware Image Synthesis
• Weakly Supervised Learning of Multi-Object 3D Scene Decompositions Using Deep Shape Priors
• A Neural Rendering Framework for Free-Viewpoint Relighting
• NeRV: Neural Representations for Videos
• ADeLA: Automatic Dense Labeling with Attention for Viewpoint Adaptation in Semantic Segmentation