Point2Primitive: CAD Reconstruction from Point Cloud by Direct Primitive Prediction

TL;DR

Point2Primitive introduces a direct primitive prediction framework for CAD reconstruction from point clouds, improving accuracy and editability over traditional implicit neural methods by explicitly predicting sketch primitives with a transformer-based decoder.

Contribution

It presents a novel approach that directly predicts explicit CAD primitives from point clouds using an improved transformer decoder, enhancing precision and topological reconstruction.

Findings

Outperforms implicit methods in primitive accuracy

Achieves higher geometric fidelity

Effectively reconstructs CAD topology via extrusion segmentation

Abstract

Recovering CAD models from point clouds requires reconstructing their topology and sketch-based extrusion primitives. A dominant paradigm for representing sketches involves implicit neural representations such as Signed Distance Fields (SDFs). However, this indirect approach inherently struggles with precision, leading to unintended curved edges and models that are difficult to edit. In this paper, we propose Point2Primitive, a framework that learns to directly predict the explicit, parametric primitives of CAD models. Our method treats sketch reconstruction as a set prediction problem, employing a improved transformer-based decoder with explicit position queries to directly detect and predict the fundamental sketch curves (i.e., type and parameter) from the point cloud. Instead of approximating a continuous field, we formulate curve parameters as explicit position queries, which are optimized autoregressively to achieve high accuracy. The overall topology is rebuilt via extrusion segmentation. Extensive experiments demonstrate that this direct prediction paradigm significantly outperforms implicit methods in both primitive accuracy and overall geometric fidelity.