NURBS-Diff: A Differentiable Programming Module for NURBS. (May 2022)

• Record Type: Journal Article
• Authors: Deva Prasad, Anjana; Balu, Aditya; Shah, Harshil; Sarkar, Soumik; Hegde, Chinmay; Krishnamurthy, Adarsh
• Is Part Of: Computer aided design. Volume 146(2022)
• Publication Date: 2022-05
• Subjects: Differentiable NURBS module -- NURBS -- Geometric deep learning -- Surface modeling; Computer-aided design -- Periodicals; Engineering design -- Data processing -- Periodicals; Computer graphics -- Periodicals; Conception technique -- Informatique -- Périodiques; Infographie -- Périodiques; Computer graphics; Engineering design -- Data processing; Periodicals; Electronic journals; 620.00420285
• DOI: 10.1016/j.cad.2022.103199
• Languages: English
• ISSNs: 0010-4485

Abstract

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. Highlights: A differentiable NURBS module for integrating splines with machine learning. Validation of the module with CAD operations such as curve and surface fitting. Integration with loss functions that impose geometric constraints in deep learning. Order of magnitude improvement over existing spline-based surface reconstruction.