Point2CAD: Reverse Engineering CAD Models from 3D Point Clouds

Computer-Aided Design (CAD) model reconstruction from point clouds is an important problem at the intersection of computer vision, graphics, and machine learning; it saves the designer significant time when iterating on in-the-wild objects. Recent advancements in this direction achieve relatively reliable semantic segmentation but still struggle to produce an adequate topology of the CAD model. In this work, we analyze the current state of the art for that ill-posed task and identify shortcomings of existing methods. We propose a hybrid analytic-neural reconstruction scheme that bridges the gap between segmented point clouds and structured CAD models and can be readily combined with different segmentation back-bones. Moreover, to power the surface fitting stage, we propose a novel implicit neural representation offreeform surfaces, driving up the performance of our overall CAD reconstruction scheme. We extensively evaluate our method on the popular ABC benchmark of CAD models and set a new state-of-the-art for that dataset. Code is available at https://github.com/YujiaLiu76Ipoint2cad.

1. Introduction

The task of reverse engineering CAD models from 3D point clouds has gained increasing attention in recent years due to the rapid development of 3D scanning technologies. Re-verse engineering involves transforming a physical object into a digital model in an editing-friendly format, which can be used for analysis, visualization, and manufacturing.