Deep Vectorization of Technical Drawings

Abstract

We present a new method for vectorization of technical line drawings, such as floor plans, architectural drawings, and 2D CAD images. Our method includes (1) a deep learning-based cleaning stage to eliminate the background and imperfections in the image and fill in missing parts, (2) a transformer-based network to estimate vector primitives, and (3) optimization procedure to obtain the final primitive configurations. We train the networks on synthetic data, renderings of vector line drawings, and manually vectorized scans of line drawings. Our method quantitatively and qualitatively outperforms a number of existing techniques on a collection of representative technical drawings.

References in corpus (8)

• Adam: A Method for Stochastic Optimization
• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• A Neural Representation of Sketch Drawings
• Learning to Infer Graphics Programs from Hand-Drawn Images
• DeepSpline: Data-Driven Reconstruction of Parametric Curves and Surfaces
• Stroke-based sketched symbol reconstruction and segmentation
• Mastering Sketching: Adversarial Augmentation for Structured Prediction
• VectorDefense: Vectorization as a Defense to Adversarial Examples

Cited by in corpus (6)

• Computer-Aided Design as Language
• 3D Parametric Wireframe Extraction Based on Distance Fields
• Structure-aware Contrastive Learning for Diagram Understanding of Multimodal Models
• Engineering Sketch Generation for Computer-Aided Design
• De-rendering Stylized Texts
• Robust reconstructions by multi-scale/irregular tangential covering