Dragonfly 3D World News

Dragonfly 官方博客介绍工业 X 射线与 CT 如何结合三维图像分析定位、测量缺陷（孔隙、裂纹、焊点），并展示 Dragonfly 3D World 2027 的实例分割（Grid-Based SAM、Cellpose-SAM）、多网格测量、STEP/CAD 偏差图、粗糙度测量和报告生成器。The official Dragonfly blog explains how industrial X-ray and CT combined with 3D image analysis locate and measure defects (porosity, cracks, solder joints), showcasing Dragonfly 3D World 2027 instance segmentation (Grid-Based SAM, Cellpose-SAM), multi-mesh measurements, STEP/CAD deviation maps, roughness measurement and Report Builder.

麦吉尔大学一篇生物考古论文的配套数据集：用 Dragonfly 3D World 2025.1 训练的神经网络对四只猫标本的 micro-CT 重建进行分割，并附 Dragonfly 团队预训练的 2.5D U-Net 模型。Supplementary dataset from McGill University for a bioarchaeology manuscript: neural networks trained in Dragonfly 3D World 2025.1 segment micro-CT reconstructions of four cat specimens, plus a pre-trained 2.5D U-Net from the Dragonfly team.