报告题目：将图像转化为知识——计算机图像分析和定量成像信息学(Turning Images into Knowledge: Computational Image Analysis and Quantitative Imaging Informatics)
报告时间：5月16日（星期三）上午10:00
报告地点：学科3号楼S410会议室
报告人：余维淼 研究员
主持人：徐 军 教授
欢迎广大师生踊跃参加！
江苏省大数据分析技术重点实验室
江苏省气象能源利用与控制工程技术研究中心
江苏省大气环境与装备技术协同创新中心
信息与控制学院
2018年5月14日
报告摘要：The past decades have seen a rapid development of different bioimaging techniques. With these hardware advancements and the generation of novel probes, we are now able to acquire 2D images or 3D videos of migrating cell in vivo or in vitro. It seems that the development of microscopy techniques is faster than our ability of interrogating those acquired images. Qualitative "seeing" is the first step towards biological discoveries, since biological images/videos contain a wealth of quantitative information, which can only be revealed with considerable computational effort. Direct application of existing image processing and computer vision techniques to new biological problems is challenging and difficult. Nowadays, biologists and computer scientists are closely working together to obtain new biological understanding, yet the complexities of biological images/videos still pose many difficult problems. Solving these problems timely facilitates complex, systematic studies for biologists, as the developed approaches can extract information that reveal previously undetected phenomena, gaining us a deeper insight.
My research focused on using computational image analysis and quantitative imaging informatics as a powerful tool to facilitate the biological studies. We are transforming modern biology from qualitative observation to a quantitative science. With improved analysis capabilities, we will be able to extract reliable quantitative knowledge from vivid images/videos. The proposed research will have an impact on different biological fields, such as cell biology, developmental biology, cancer studies, stem cell research, drug discovery and mechanobiology. In this sharing session, I will share how computational image analysis can help us to better understanding of different biological projects. We will first start with an accurate cell segmentation approach, Evolving Generalized Voronoi Diagram. Then few projects will be briefly presented, including Quantitative Understanding of Cell Protrusion during Cell Migration, Collective Cell Migration in Drosophila Oocyte, Reconstruction of Zebrafish Embryo, etc. The concept of 3D histology stack reconstruction will also be briefly discussed. I wish I have time to cover some other clinical relevant studies at the end of the talk.
关键词:
Computational Image Analysis; Quantitative Imaging Informatics; mage Segmentation(2-D/3-D); Computational Morphodynamics (Morphological Dynamics); Cell Migration and Tracking .
报告人简介: 余维淼博士2007年毕业于新加坡国立大学（National University of Singapore）数字图像处理和机器视觉专业。他在新加坡生物信息研究院（Bioinformatics Institute）图像信息部从事了三年半的博士后（Postdoc Fellow）研究工作，此后于2010年他作为高级博士后（Senior Postdoc Fellow）加入了分子细胞研究院（Institute of Molecular and Cell Biology）光学成像中心，并于2012年成为独立研究员（Research Scientist）负责该研究院的生物图像处理与信息学的研究方向。如今他作为分子细胞研究院该学科主要负责人（Head of Computational Bioimage Analysis Unit/Senior Scientist），主导生物图像计算分析的相关研究与开发工作。他的主要科研成果发表在顶级科技期刊，如Current Biology, BMC Biology, Plos Generic, Bioinformatics, Nature Material, Nature Communication以及 Nature Cell Biology等等。
余维淼博士的研究方向是以数字计算为基础的生物医学图像分析和定量的图像信息学。该方向指的是用计算方法实现2维至4维的生物图像的自动处理并从中提取定量的相关信息。该方向以开发相应的分析算法和软件来定量分析不同条件下细胞的形态与运动、生物组织的形态、蛋白质的形态与分布、以及其他相关的生物信息。这些信息对于更好的从生物学和生物物理学角度来理解复杂的细胞至关重要。近五年来，他和多个药厂和医院建立的紧密的合作关系，在病理图像、胶原蛋白成像和免疫细胞成像的图像数据分析等方向做了大量扎实的研究工作。他的研究方法包含了多个传统学科，其中包括信号处理、图像处理和计算机视觉、优化、机器学习、模式识别和数学建模等等。他的研究成果在药品开发以及临床诊断产生了广泛的经济效益。