Jhorse/BrickNet

BrickNet数据集是一个砖墙图案的集合，专为机器学习、计算设计和机器人制造交叉领域的研究而开发。该数据集支持建筑美学变化、模式多样性以及机器人干堆墙图案的算法生成研究。研究应用包括：生成建模（训练变分自编码器VAEs、GANs或扩散模型用于墙设计生成）、模式识别（聚类和分析结构/美学设计趋势）、制造研究（机器人组装、结构优化和材料驱动设计研究）以及创意探索（计算建筑和艺术中的设计空间探索）。数据集包含30,000多个独特的墙样本，每个墙有180多个特征，文件格式为CSV（行代表墙，列代表砖/特征），可以重塑为网格（20×9）。领域涉及建筑几何、材料系统和制造感知设计。

The BrickNet Dataset is a collection of brick wall patterns developed for research at the intersection of machine learning, computational design, and robotic fabrication. The dataset supports architectural studies of aesthetic variation, pattern diversity, and algorithmic generation of wall patterns for robotic dry stacking. Research Applications include: Generative Modeling: Training Variational Autoencoders (VAEs), GANs, or Diffusion Models for wall design generation; Pattern Recognition: Clustering and analyzing structural/aesthetic design tendencies; Fabrication Research: Robotic assembly, structural optimization, and material-driven design studies; Creative Exploration: Design space exploration in computational architecture and art. Dataset Overview: Dataset Name: BrickNet; File Format: CSV (rows = walls, columns = bricks/features); Size: 30,000+ unique wall samples, 180+ features per wall; Reshaping: Rows can be reshaped into grids (20×9); Domain: Architectural geometry, material systems, fabrication-aware design.