NURBS-Diff: A Differentiable Programming Module for NURBS

TL;DR

This paper introduces a differentiable NURBS module compatible with deep learning frameworks, enabling CAD model integration, surface fitting, and point cloud reconstruction with improved performance.

Contribution

We develop a GPU-accelerated, differentiable NURBS module that can be integrated into deep learning frameworks like PyTorch for CAD and geometric tasks.

Findings

Effective in CAD operations like surface offsetting

Improves deep learning tasks such as point cloud reconstruction

Compatible with GPU acceleration and existing frameworks

Abstract

Boundary representations (B-reps) using Non-Uniform Rational B-splines (NURBS) are the de facto standard used in CAD, but their utility in deep learning-based approaches is not well researched. We propose a differentiable NURBS module to integrate NURBS representations of CAD models with deep learning methods. We mathematically define the derivatives of the NURBS curves or surfaces with respect to the input parameters (control points, weights, and the knot vector). These derivatives are used to define an approximate Jacobian used for performing the "backward" evaluation to train the deep learning models. We have implemented our NURBS module using GPU-accelerated algorithms and integrated it with PyTorch, a popular deep learning framework. We demonstrate the efficacy of our NURBS module in performing CAD operations such as curve or surface fitting and surface offsetting. Further, we show its utility in deep learning for unsupervised point cloud reconstruction and enforce analysis constraints. These examples show that our module performs better for certain deep learning frameworks and can be directly integrated with any deep-learning framework requiring NURBS.