View Synthesis by Appearance Flow

TL;DR

This paper introduces a CNN-based method for novel view synthesis that predicts appearance flows to copy pixels from input images, enabling high-quality rendering of new viewpoints for objects and scenes.

Contribution

It proposes a novel appearance flow prediction framework that improves view synthesis by explicitly learning pixel correspondences, generalizing to multiple input views.

Findings

Synthesizes higher perceptual quality images than previous CNN methods.

Effectively generalizes to multiple input views.

Applicable to both objects and scenes.

Abstract

We address the problem of novel view synthesis: given an input image, synthesizing new images of the same object or scene observed from arbitrary viewpoints. We approach this as a learning task but, critically, instead of learning to synthesize pixels from scratch, we learn to copy them from the input image. Our approach exploits the observation that the visual appearance of different views of the same instance is highly correlated, and such correlation could be explicitly learned by training a convolutional neural network (CNN) to predict appearance flows -- 2-D coordinate vectors specifying which pixels in the input view could be used to reconstruct the target view. Furthermore, the proposed framework easily generalizes to multiple input views by learning how to optimally combine single-view predictions. We show that for both objects and scenes, our approach is able to synthesize novel views of higher perceptual quality than previous CNN-based techniques.