LS-CAD
收藏资源简介:
LS-CAD是由上海科技大学和DGene数字技术共同创建的一个大规模CAD数据集,专门用于支持建筑CAD图纸中的符号识别研究。该数据集包含50个标注的楼层平面图,每个平面图的平均覆盖面积为1000平方米,远超现有数据集的规模。数据集的创建过程遵循精细的标注标准,确保了符号识别任务的高质量基准。LS-CAD数据集的应用领域主要集中在建筑CAD图纸的自动化处理,旨在提高3D室内建模和代码合规性检查等任务的效率。
LS-CAD is a large-scale CAD dataset jointly created by ShanghaiTech University and DGene Digital Technology, which is specifically designed to support symbol recognition research in architectural CAD drawings. This dataset contains 50 annotated floor plans, with an average covered area of 1,000 square meters per plan, far exceeding the scale of existing datasets. The dataset was developed following rigorous annotation standards to ensure a high-quality benchmark for symbol recognition tasks. The application fields of the LS-CAD dataset mainly focus on the automated processing of architectural CAD drawings, aiming to improve the efficiency of tasks such as 3D indoor modeling and code compliance checking.

- 1CADSpotting: Robust Panoptic Symbol Spotting on Large-Scale CAD Drawings上海科技大学 · 2024年



