TransCAD: A Hierarchical Transformer for CAD Sequence Inference from Point Clouds

TL;DR

TransCAD is a hierarchical transformer model that infers CAD sequences from point clouds, advancing 3D reverse engineering with state-of-the-art accuracy and a new evaluation metric.

Contribution

It introduces TransCAD, a novel hierarchical transformer architecture with a loop refiner for CAD sequence inference from point clouds, outperforming existing methods.

Findings

Achieves state-of-the-art results on DeepCAD and Fusion360 datasets.

Introduces the mean Average Precision of CAD Sequence metric.

Demonstrates effective hierarchical learning for CAD sequence prediction.

Abstract

3D reverse engineering, in which a CAD model is inferred given a 3D scan of a physical object, is a research direction that offers many promising practical applications. This paper proposes TransCAD, an end-to-end transformer-based architecture that predicts the CAD sequence from a point cloud. TransCAD leverages the structure of CAD sequences by using a hierarchical learning strategy. A loop refiner is also introduced to regress sketch primitive parameters. Rigorous experimentation on the DeepCAD and Fusion360 datasets show that TransCAD achieves state-of-the-art results. The result analysis is supported with a proposed metric for CAD sequence, the mean Average Precision of CAD Sequence, that addresses the limitations of existing metrics.