Data underlying the MSc thesis: Deep learning segmentation of 3D ultrasound thyroid imaging

3D ultrasound data acquired with a matrix probe and philips system in february 2023.Dataset_PHILIPS_save3D_and_saveMPR:Left and right thyroid lobe scan of 57 healthy volunteers.Data is ordered by participant numbers, containing 10 files per patient:3 MPR files (axial, coronal and saggital orientation) per lobe (higher resolution)1 Save3D DICOM (philips) file per lobe (more slices)This data was not altered after collection from the US system.<br>Annotated 27 subjects:Contains the first 27 subjects of the folder Dataset_PHILIPS_save3D_and_saveMPR_35GB. Philips DICOM tags were changed to regular DICOM tags and samples were rotated. Contains one annotation file per subject containing thyroid, jugular vein and cartid artery.<br>More information can be found in my thesis: deep learning segmentation of 3D ultrasound thyroid imaging

本数据集为2023年2月使用矩阵探头与飞利浦（Philips）系统采集的三维超声数据。Dataset_PHILIPS_save3D_and_saveMPR：包含57名健康志愿者的左右甲状腺叶扫描数据。数据按受试者编号排序，每名受试者对应10个文件：每个甲状腺叶对应3个MPR（Multi-Planar Reconstruction，多平面重建）文件，分别为轴位、冠状位与矢状位，分辨率更高；每个甲状腺叶对应1个Save3D格式的飞利浦DICOM文件，包含更多断层切片。该数据从超声系统采集后未经过任何修改。
标注子集（共27名受试者）：包含Dataset_PHILIPS_save3D_and_saveMPR_35GB文件夹中的前27名受试者数据。已将飞利浦专属DICOM标签转换为标准DICOM标签，并对样本进行了旋转处理。每名受试者对应1个标注文件，标注内容涵盖甲状腺、颈静脉与颈动脉。
更多详细信息可参阅我的学位论文：《三维超声甲状腺影像的深度学习分割》