Z2P: Instant Visualization of Point Clouds

TL;DR

This paper introduces Z2P, a neural network-based method for instant visualization of point clouds as surface-like images, bypassing complex reconstruction and normal estimation, with optional control over appearance.

Contribution

Z2P is a novel neural network approach that directly translates point cloud depth maps into surface-like images for rapid visualization.

Findings

Instant visualization of point clouds achieved.

Handles noise and non-uniform sampling effectively.

Supports appearance conditioning like color and lighting.

Abstract

We present a technique for visualizing point clouds using a neural network. Our technique allows for an instant preview of any point cloud, and bypasses the notoriously difficult surface reconstruction problem or the need to estimate oriented normals for splat-based rendering. We cast the preview problem as a conditional image-to-image translation task, and design a neural network that translates point depth-map directly into an image, where the point cloud is visualized as though a surface was reconstructed from it. Furthermore, the resulting appearance of the visualized point cloud can be, optionally, conditioned on simple control variables (e.g., color and light). We demonstrate that our technique instantly produces plausible images, and can, on-the-fly effectively handle noise, non-uniform sampling, and thin surfaces sheets.