arXiv:2511.14870v1 Announce Type: new 
Abstract: This paper presents a novel geometric representation for CAD Boundary Representation (B-Rep) based on volumetric distance functions, dubbed B-Rep Distance Functions (BR-DF). BR-DF encodes the surface mesh geometry of a CAD model as signed distance function (SDF). B-Rep vertices, edges, faces and their topology information are encoded as per-face unsigned distance functions (UDFs). An extension of the Marching Cubes algorithm converts BR-DF directly into watertight CAD B-Rep model (strictly speaking a faceted B-Rep model). A surprising characteristic of BR-DF is that this conversion process never fails. Leveraging the volumetric nature of BR-DF, we propose a multi-branch latent diffusion with 3D U-Net backbone for jointly generating the SDF and per-face UDFs of a BR-DF model. Our approach achieves comparable CAD generation performance against SOTA methods while reaching the unprecedented 100% success rate in producing (faceted) B-Rep models.

تقدم هذه الورقة وظائف المسافة B-Rep (BR-DF)، وهي تمثيل هندسي جديد لتمثيل الحدود CAD (B-Rep) يعتمد على وظائف المسافة الحجمية. تقوم BR-DF بتشفير هندسة شبكة السطح لنموذج CAD كوظائف مسافة موقعة (SDF) وتستخدم وظائف المسافة غير الموقعة (UDF) لكل وجه لرؤوس وحواف ووجوه B-Rep. يتيح تمديد خوارزمية Marching Cubes تحويل BR-DF مباشرة إلى نموذج B-Rep CAD محكم، محققًا معدل نجاح بنسبة 100% في الإنتاج.

El artículo presenta las B-Rep Distance Functions (BR-DF), una nueva representación geométrica para la Representación de Frontera CAD (B-Rep) basada en funciones de distancia volumétricas. Las BR-DF codifican la geometría de la malla de superficie de un modelo CAD como funciones de distancia firmadas (SDF) y utilizan funciones de distancia no firmadas (UDF) por cara para los vértices, bordes y caras. Una extensión del algoritmo Marching Cubes convierte BR-DF directamente en modelos B-Rep CAD estancos, logrando una tasa de éxito del 100% en la generación.

Cet article présente les B-Rep Distance Functions (BR-DF), une nouvelle représentation géométrique pour la représentation de frontière CAD (B-Rep) basée sur des fonctions de distance volumétriques. Les BR-DF codent la géométrie de maillage de surface d'un modèle CAD en tant que fonctions de distance signées (SDF) et utilisent des fonctions de distance non signées (UDF) par face pour les sommets, arêtes et faces. Une extension de l'algorithme Marching Cubes permet la conversion des BR-DF en modèles B-Rep CAD étanches, atteignant un taux de réussite de 100 % dans la génération.

The paper introduces B-Rep Distance Functions (BR-DF), a new geometric representation for CAD Boundary Representation (B-Rep) using volumetric distance functions. BR-DF encodes CAD model surface mesh geometry as signed distance functions (SDF) and utilizes per-face unsigned distance functions (UDFs) for vertices, edges, and faces. An extension of the Marching Cubes algorithm enables the conversion of BR-DF into watertight CAD B-Rep models, achieving a 100% success rate in generation.

B-Rep Distance Functions (BR-DF): How to Represent a B-Rep Model by Volumetric Distance Functions?

arXiv:2512.03018v1 Announce Type: new 
Abstract: The boundary representation (B-Rep) is the standard data structure used in Computer-Aided Design (CAD) for defining solid models. Despite recent progress, directly generating B-Reps end-to-end with precise geometry and watertight topology remains a challenge. This paper presents AutoBrep, a novel Transformer model that autoregressively generates B-Reps with high quality and validity. AutoBrep employs a unified tokenization scheme that encodes both geometric and topological characteristics of a B-Rep model as a sequence of discrete tokens. Geometric primitives (i.e., surfaces and curves) are encoded as latent geometry tokens, and their structural relationships are defined as special topological reference tokens. Sequence order in AutoBrep naturally follows a breadth first traversal of the B-Rep face adjacency graph. At inference time, neighboring faces and edges along with their topological structure are progressively generated. Extensive experiments demonstrate the advantages of our unified representation when coupled with next-token prediction for B-Rep generation. AutoBrep outperforms baselines with better quality and watertightness. It is also highly scalable to complex solids with good fidelity and inference speed. We further show that autocompleting B-Reps is natively supported through our unified tokenization, enabling user-controllable CAD generation with minimal changes. Code is available at https://github.com/AutodeskAILab/AutoBrep.

تم تقديم نموذج جديد يسمى AutoBrep، وهو نموذج Transformer يقوم بتوليد تمثيلات الحدود (B-Reps) في التصميم المدعوم بالحاسوب (CAD) بجودة عالية وتوبولوجيا صالحة. يتناول هذا النموذج التحدي المتمثل في توليد B-Reps من البداية إلى النهاية من خلال استخدام مخطط توكين موحد يقوم بتشفير الخصائص الهندسية والتوبولوجية كرموز منفصلة، مما يسهل عملية الاستدلال عبر رسم بياني لجوار الوجوه في B-Rep.

Se ha presentado un nuevo modelo Transformer llamado AutoBrep, que genera representaciones de frontera (B-Reps) en el diseño asistido por computadora (CAD) con alta calidad y topología válida. Este modelo aborda el desafío de la generación de B-Reps de extremo a extremo mediante un esquema de tokenización unificado que codifica características geométricas y topológicas como tokens discretos, facilitando un recorrido en amplitud del grafo de adyacencia de caras de B-Rep durante la inferencia.

Un nouveau modèle Transformer nommé AutoBrep a été introduit pour générer des représentations de frontière (B-Reps) dans la conception assistée par ordinateur (CAO) avec une qualité élevée et une topologie valide. Ce modèle répond au défi de la génération de B-Reps de bout en bout en utilisant un schéma de tokenisation unifié qui encode les caractéristiques géométriques et topologiques en tant que tokens discrets, facilitant un parcours en largeur du graphe d'adjacence des faces de B-Rep lors de l'inférence.

A novel Transformer model named AutoBrep has been introduced to generate boundary representations (B-Reps) in Computer-Aided Design (CAD) with high quality and valid topology. This model addresses the challenge of end-to-end generation of B-Reps by employing a unified tokenization scheme that encodes geometric and topological characteristics as discrete tokens, facilitating a breadth-first traversal of the B-Rep face adjacency graph during inference.

B-Rep Distance Functions (BR-DF): How to Represent a B-Rep Model by Volumetric Distance Functions?

Was this article worth reading? Share it

LucidQuery AI

VSDECO

Z3D