Ocular Prosthesis Point Cloud Dataset for Patient-Specific Surface Modeling

This record contains a de-identified ocular prosthesis point cloud dataset used in the study “Riemannian Feature-Based Point Cloud Learning for Patient-Specific Ocular Prosthesis Surface Modeling.” The dataset consists of processed three-dimensional point clouds of ocular prostheses collected in collaboration with a partner hospital. The data have been processed and de-identified for research use. In addition to the point cloud data, this record also provides the RFG-trained model records for each ocular prosthesis sample. These model records were obtained after training the Riemannian Feature Generator (RFG) and can be used to compute point-wise Riemannian metric features for the corresponding ocular prosthesis point clouds. They are provided to support reproducibility of the Riemannian feature extraction and local surface reconstruction experiments reported in the manuscript.本数据集为经去标识化处理的义眼（ocular prosthesis）点云数据集，用于研究《基于黎曼特征的点云学习与个性化义眼表面建模》（Riemannian Feature-Based Point Cloud Learning for Patient-Specific Ocular Prosthesis Surface Modeling）。本数据集包含经处理的义眼三维点云数据，由我方与合作医院联合采集，所有数据均已完成处理与去标识化操作，仅供科研使用。除上述点云数据外，本数据集还提供了针对每个义眼样本的经RFG训练的模型记录。此类模型记录通过训练黎曼特征生成器（Riemannian Feature Generator, RFG）后得到，可用于为对应义眼点云逐点计算黎曼度量特征。本数据集提供此类记录，旨在支持论文中报道的黎曼特征提取与局部表面重建实验的可复现性。