目前我没有选择复现这篇论文了,直接使用最新的Transformer技术,但是借鉴了论文中提到的重建损失和空间关系损失,以及添加了一些其它的角度位移构建的损失,效果自我感觉比论文提到的网络结构强,但是这篇论文和Tooth_Alignment_Network_Based_on_Landmark_Constraints_and_Hierarchical_Graph_Structure都没提到牙齿碰撞问题,牙齿之...
正畸学专注于矫正牙齿，以改善其功能与美观。6D位姿预测涉及预测每颗牙齿在三维空间中的平移和旋转，合计六个自由度。目标：输入原始的牙齿排列，输出理想排列方案，采用监督学习。一. FDI标记法与Tooth Centering 采用FDI标记法仅保留牙冠，以每颗牙的几何中心为原点，调整点云内其他点坐标，集合C包含所有...
Tooth arrangement6D pose predictionStructureGraph neural networkDetermining optimal target tooth arrangements is a key step of treatment planning in digital orthodontics. Existing practice for specifying the target tooth arrangement involves tedious manual......
Determining optimal target tooth arrangements is a key step of treatment planning in digital orthodontics. Existing practice for specifying the target tooth arrangement involves tedious manual operations with the outcome quality depending heavily on the experience of individual specialists, leading to ineffici...