48,434 skeletal radiographs
收藏资源简介:
该数据集由慕尼黑工业大学的医院PACS系统导出,包含48,434张骨骼X光片,每张图像都标注了14个常见的解剖区域。数据集的大小和标签的准确性对于医学影像分析至关重要,尤其是在低标签资源的情况下。数据集的创建过程包括从PACS系统中提取DICOM文件,并进行数据增强和清理,以提高模型的预测准确性。该数据集主要用于解剖区域分类,旨在提高医学影像数据的质量,并为未来的医学研究和机器学习应用提供高质量的数据支持。
This dataset is exported from the hospital PACS system of the Technical University of Munich, consisting of 48,434 skeletal X-ray images. Each image is annotated with 14 common anatomical regions. The size of the dataset and the accuracy of its labels are critical for medical image analysis, particularly in low-resource annotation scenarios. The dataset creation workflow includes extracting DICOM files from the PACS system, followed by data augmentation and cleaning to enhance the prediction accuracy of machine learning models. This dataset is primarily intended for anatomical region classification, with the goal of improving the quality of medical image data and providing high-quality data support for future medical research and machine learning applications.

- 1Self-Supervised Radiograph Anatomical Region Classification -- How Clean Is Your Real-World Data?慕尼黑工业大学 · 2024年



