# Neural Point-Based Graphics

## TL;DR

This paper introduces a neural point-based graphics method that models scene appearance directly from raw point clouds with learnable descriptors, enabling photorealistic rendering without explicit surface reconstruction.

## Contribution

It proposes a novel point-based scene representation with neural descriptors and a deep rendering network, avoiding surface estimation and meshing for complex scene modeling.

## Key findings

- Effective for complex scenes from RGB-D and RGB data
- Produces photorealistic views from new viewpoints
- Handles challenging objects better than mesh-based methods

## Abstract

We present a new point-based approach for modeling the appearance of real scenes. The approach uses a raw point cloud as the geometric representation of a scene, and augments each point with a learnable neural descriptor that encodes local geometry and appearance. A deep rendering network is learned in parallel with the descriptors, so that new views of the scene can be obtained by passing the rasterizations of a point cloud from new viewpoints through this network. The input rasterizations use the learned descriptors as point pseudo-colors. We show that the proposed approach can be used for modeling complex scenes and obtaining their photorealistic views, while avoiding explicit surface estimation and meshing. In particular, compelling results are obtained for scene scanned using hand-held commodity RGB-D sensors as well as standard RGB cameras even in the presence of objects that are challenging for standard mesh-based modeling.