FoV-Net: Rotation-Invariant CAD B-rep Learning via Field-of-View Ray Casting

2026-02-27

AI Summary

This work proposes the first fully rotation-invariant learning framework for boundary representations (B-reps). Existing methods rely on absolute coordinates and surface normals, rendering them sensitive to SO(3) rotations and limiting their generalization. To address this, the proposed approach constructs UV grids using local reference frames (LRFs) and encodes global structural context via field-of-view (FoV) ray casting, thereby achieving rotation-invariant representations that preserve fine-grained geometric details. Features are then propagated across the B-rep graph using a lightweight CNN combined with a graph attention network (GAT). The method achieves state-of-the-art performance on B-rep classification and segmentation tasks, demonstrating high robustness to arbitrary rotations and significantly improved generalization and data efficiency—even when trained with substantially less data.

Technology Category

Computer Vision: Representation Learning for VisionMachine Learning: Representation LearningKnowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphsSearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingSemantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

Abstract

Learning directly from boundary representations (B-reps) has significantly advanced 3D CAD analysis. However, state-of-the-art B-rep learning methods rely on absolute coordinates and normals to encode global context, making them highly sensitive to rotations. Our experiments reveal that models achieving over 95% accuracy on aligned benchmarks can collapse to as low as 10% under arbitrary $\mathbf{SO}(3)$ rotations. To address this, we introduce FoV-Net, the first B-rep learning framework that captures both local surface geometry and global structural context in a rotation-invariant manner. Each face is represented by a Local Reference Frame (LRF) UV-grid that encodes its local surface geometry, and by Field-of-View (FoV) grids that capture the surrounding 3D context by casting rays and recording intersections with neighboring faces. Lightweight CNNs extract per-face features, which are propagated over the B-rep graph using a graph attention network. FoV-Net achieves state-of-the-art performance on B-rep classification and segmentation benchmarks, demonstrating robustness to arbitrary rotations while also requiring less training data to achieve strong results.

Problem

• rotation-invariance
• B-rep learning
• 3D CAD
• SO(3) sensitivity
• geometric deep learning

Innovation

• rotation-invariant
• B-rep learning
• Field-of-View ray casting
• Local Reference Frame
• graph attention network