Dr. Qiong Xu, a former member of GSQ, has joined Chinese Academy of Sciences!
I’m Qiong XU. I studied as a joint researcher of Radiation Oncology Department of Stanford University and GI-CoRE GSQ of Hokkaido University from Nov. 2016 to Mar. 2018. During my time in GSQ, my research mainly focuses on X-ray CT imaging, including reconstruction algorithm and data correction method development. I proposed a region-specific strategy for dictionary learning-based CT reconstruction method. With the regional consideration, the heterogeneity of the structure and noise inside of the object is well balanced. This method performed very well for low-dose thoracic CT. The journal paper of this work has been submitted to IEEE Transactions on Medical Imaging and under the first round revision now. Another topic I studied is metal artifact reduction. Metal artifacts have severely influenced the accuracy of radiation diagnosis and therapy. I developed a deep learning-based metal artifact reduction method, which used the deep image prior method to restore the projection data crossing the metals. This proposed method was very effective from our primary simulation study. The preliminary results have been accepted as oral presentation in 2018 AAPM annual meeting. In addition, I also put some efforts on supervising the PhD student - Yusuke Nomura’s research. We improved the scatter-kernel-based scatter correction method and studied the scatter corrections effect for proton therapy. The journal paper of this work has been submitted into Physics in Medicine and Biology. Besides that, we also developed a deep learning method for scatter correction. A residual convolutional neural network was utilized to estimate the scatter. This work has also been accepted as oral presentation in 2018 AAPM annual meeting. Now, we are still trying to improve this work. After I finished my study in Hokkaido University, I joined Institute of High Energy Physics Chinese Academy of Sciences as an associate professor. I still work on X-ray imaging field. Right now, I mainly work on the development of spectral CT reconstruction method. To improve the CNR and decrease the dose of our spectral CT system from the reconstruction algorithm aspect is my goal. Qiong Xu, Ph.D. Institute of High Energy Physics Chinese Academy of Sciences