Architecture and Building the Medical Image Anonymization Service: Cloud, Big Data and Automation

Medical images provide significant information to assist patients to obtain correct treatments. In order to create more innovative medical applications, NCHC with the three major medical institutions builds Taiwan's first medical image annotation database. However, under the restriction of personal information law in Taiwan, data privacy is a critical issue. Besides, the data from institutions are a huge amount to process. Thus, we propose “Anonymization Cloud Service” (ACS), based on cloud computing and big data architecture to serve privacy data to be anonymized with DICOM or CSV format. The research has three main components: (1) data conversion between DICOM and CSV, and transmission between transform node and big data cluster, (2) parallelized anonymous process with automation API, and (3) privacy control mechanism for risk evaluation. The results are implemented in more than tens of thousands of cases, over 10 TB of data, and 10 million DICOM files. Those actual cases show the proposed method effectively and efficiently achieves the goals of high performance and low risk.