Auto-Regressive Surface Cutting

2025-06-22

AI Summary

Existing 3D mesh surface cutting methods often yield semantically incoherent cuts and geometric fragmentation. To address this, we propose SeamGPT—the first approach to formulate seam generation as an autoregressive sequence prediction task. SeamGPT jointly encodes mesh vertices and edges via point cloud sampling and incorporates shape-conditioned embeddings into a GPT-style Transformer architecture; seams are represented as quantized 3D coordinates, enabling unified handling of both manifold and non-manifold meshes. Our method significantly improves semantic coherence and structural integrity of seams, achieving state-of-the-art performance on UV-unwrapping benchmarks. It supports interactive artist workflows and post-processing of 3D-scanned models, while also producing high-quality boundaries for part segmentation—thereby enhancing the robustness and practicality of downstream 3D segmentation tools.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Surface cutting is a fundamental task in computer graphics, with applications in UV parameterization, texture mapping, and mesh decomposition. However, existing methods often produce technically valid but overly fragmented atlases that lack semantic coherence. We introduce SeamGPT, an auto-regressive model that generates cutting seams by mimicking professional workflows. Our key technical innovation lies in formulating surface cutting as a next token prediction task: sample point clouds on mesh vertices and edges, encode them as shape conditions, and employ a GPT-style transformer to sequentially predict seam segments with quantized 3D coordinates. Our approach achieves exceptional performance on UV unwrapping benchmarks containing both manifold and non-manifold meshes, including artist-created, and 3D-scanned models. In addition, it enhances existing 3D segmentation tools by providing clean boundaries for part decomposition.

Problem

• Generates coherent cutting seams for 3D meshes
• Improves UV unwrapping and texture mapping quality
• Enhances 3D segmentation with clean part boundaries

Innovation

• Auto-regressive model for seam generation
• Next token prediction for surface cutting
• GPT-style transformer for 3D coordinates