报告人：王森 博士 斯坦福大学医学院放射学系
主持人：陈阳
报告时间：2024年12月25日（周三）下午16:00
报告地点：东南大学九龙湖校区计算机楼513报告厅
报告摘要：Photon Counting Detectors (PCDs) for X-ray Computed Tomography provide a semiconductor-based energy-differentiating counting mechanism that better captures realistic physics processes. The direct impacts include significant image resolution improvements and dose reduction compared to conventional energy integrating detectors (EIDs), which has been widely evaluated and recognized. Apart from that, in this presentation, we aim to discuss how the accurate physics information from PCDs potentially offers more extensive benefits for X-ray imaging. We have found that the physics information acquired by PCDs strongly synergizes with current machine learning approaches. By incorporating this physics information, we were able to enable a self-supervised denoising method (Noise2Noise) and an end-to-end differentiable Photon Counting CT (PCCT) imaging chain. These preliminary trials can be generally applied to other imaging methods or applications that involve count measurements and optimizations (in inverse problems).
报告人简介：王森，2014年本科毕业于清华大学工程物理系获学士学位，并于2019年在清华大学工程物理系获博士学位（导师：张丽研究员）。博士毕业同年获清华大学“紫荆学者”计划资助，于2020年赴斯坦福放射学系（Department of Radiology）从事博士后研究，合作导师为：Adam Wang教授和Norbert Pelc教授，并于2023年转为研究科学家（Research Scientist）。他的研究方向主要为先进X射线能谱成像在医学诊断和介入手术中的应用，代表性工作包括先进光子计数CT探测器物理分析、准单能光子计数CT成像、物理信息约束自监督AI低剂量CT成像等，近期研究方向: 1）低剂量CT成像与AI扫描控制，2）数据驱动的高质量定量X射线成像及其全流程物理信息微分分析。
