A Multifunctional Image Processing Tool for CT Data Standardization

As one of the most commonly used imaging examinations in clinical practice, CT images have been widely sued in computer-aided diagnosis and disease-related studies. Due to the diverse content and format of the original CT data, it is usually necessary to conduct standardized operations on CT data before image processing or deep learning, which is called data preprocessing. The complexity of CT data makes preprocessing often time-consuming and labor-intensive especially for multi-cohort studies of multiple sequence CT images, and the results directly affect the subsequent algorithms. This study establishes a universal CT data standardization process and tool that can greatly improve the efficiency of preprocessing. The established pipeline includes de-privacy, format conversion, data check, denoising, resampling and registration. The standardization process we built in the tool is applicable to the preprocessing of most studies. We integrate a variety of common algorithms to extend the user’s choice at the same time.