LAUSR

Metal artifacts seriously degrade the quality of the CT data and bring great difficulties to subsequent image processing and analysis, which nowadays become a great concern in X-ray CT applications. In this paper, we introduce a U-net-based metal artifact reduction method into CT image domain. The proposed network reduces metal artifacts by learning an end-to-end mapping of images from metal-corrupted CT images to their corresponding artifact-free ground truth images. We design and optimize the network through the simulation experiments. The experimental results show that the proposed method can well reduce metal artifacts of CT images, and this method has higher computational efficiency and greatly shortens the processing time. It avoids complex image preprocessing and can accept input images of any size. Therefore, it can be a more automated way to handle large amounts of data, making it ideal for existing CT workflows.