第6回計算解剖学セミナー

ご参加ありがとうございました.当日の写真は ココ をクリックしてください

開催概要

プログラム

 講師 : Tobias Heimann, PhD (German Cancer Research Center, Germany)
th.jpg
  • タイトル
    • Statistical shape models for the automatic segmentation of 3D medical image data
  • 概要
    • Statistical models are a powerful tool for the automatic analysis of medical image data. Due to their inherent learning capabilities, they offer a general framework to tackle a variety of different problems. This talk will present the basic concepts and techniques to design, build, and employ statistical models of shape and appearance. Challenges and possible solutions on the way to produce robust models will be pointed out. A selection of applications based on different imaging modalities and structures of interest will show strengths and difficulties of model-based segmentation. Finally, current trends and challenges in the field will be discussed.
  • 略歴
    • Dr. Tobias Heimann studied Medical Informatics at the University of Heidelberg and received a PhD in Computer Science for his work at the German Cancer Research Center in 2008. After his PhD studies, he worked as a European Marie-Curie fellow in the Asclepios team at INRIA Sophia Antipolis, France. Since 2010, he is working again at the German Cancer Research Center in the Division of Medical and Biological Informatics. Dr. Heimann is a referee for several international medical imaging journals and conferences. In 2007, he initiated the MICCAI Grand Challenge workshops for the evaluation of medical image analysis algorithms on clinical data. Until now, he has published over 40 peer-reviewed journal and conference articles.
  • Homepage


 講師 : Hans Lamecker, PhD (Zuse Institute Berlin, Germany)
HansLamecker.png
  • タイトル
    • Modeling Anatomical Shapes: Matching and Deforming Meshes
  • 概要
    • The reconstruction of geometrical shape from 3d medical image data is a challenging problem. Nevertheless, it is a fundamental prerequisite in patient-specific therapy planning, e.g. using finite-element simulations. Here, shapes are usually represented as meshes. Such applications generally impose two fundamental requirements on the properties of meshes: (1) true representation of complex anatomy, (2) high-quality and low number of elements. Furthermore, automatic solutions are generally desired. A promising path for solving the reconstruction problem are model-based approaches. Here, a-priori knowledge about shape and other application-specific characteristics is incorporated into the algorithm. Generating such models requires (a) to match shapes, while adapting the models to image data needs (b) suitable deformation models. I will present solutions for both problems that satisfy the application-specific requirements (1) and (2), and discuss their strengths and limitations.
  • 略歴
    • Dr. Hans Lamecker graduated in physics from the University of Heidelberg, Germany. He performed parts of his studies at the University of Auckland, NZ, and at the MIT in Cambridge, USA. Since 2000 he works as a research assistant at the Zuse Institute Berlin, Germany. He received a PhD in mathematics and computer science from the Free University in Berlin in 2008. In 2009, he worked as a research assistant in the Asclepios Research Group at INRIA Sophia Antipolis, France. Dr. Lamecker's research interests lie in model-based geometry reconstruction from 3d image data, with a focus on solving application-driven problems in bio-medicine.
  • Homepage


 講師 : Marius George Linguraru, PhD (National Institutes of Health, USA)
Marius3BW.png
  • タイトル : Multi-organ modeling and analysis in abdominal CT
  • 概要
    • In computed-tomography (CT)-based clinical abdominal diagnosis, radiologists analyze multi-phase CT data. This is typically done organ-by-organ and slice-by-slice until the entire image data are covered. While computer-aided diagnosis (CAD) and medical image analysis traditionally focus on organ- or disease-based applications, there is a strong incentive to migrate toward the automated simultaneous segmentation and analysis of multiple organs for comprehensive diagnosis or pre-operative planning and guidance. Challenges for the automated abdominal organ analysis include the large shape variability and the similar appearance and close proximity of different organs. Nevertheless, the interpretation of medical images benefits from anatomical and physiological priors to optimize CAD applications, such as location and enhancement. This allows modeling the abdomen using landmarks, location probability maps, and pose distributions. Segmentation methods can be initialized from abdominal models and then refined using patient specific information. An example is presented using a new formulation of a 4D directional graph to automatically segment the liver, spleen, and left and right kidneys. A maximum a posteori probability framework to determine the pose of abdominal organs, including the pancreas, is also discussed. The approach promises to support the processing of large medical data in a clinically-oriented holistic analysis of the abdomen.
  • 略歴
    • Marius George Linguraru received a D.Phil. in Medical Image Analysis from the University of Oxford, UK. His earlier studies comprise an M.A. in Cultural Studies, an M.S. in Processing Systems, and a B.S. in Computer Science from the University of Sibiu, Romania. His training includes a postdoctoral fellowship at Harvard University in Cambridge, MA.
      Since 2007, he is Staff Scientist in the Department of Radiology and Imaging Sciences at the Clinical Center, National Institutes of Health (NIH), Bethesda, MD. Previously, he was Expert Engineer in the Asclepios Research Group at the National Institute of Research in Computer Science (INRIA) in Sophia Antipolis, France, and Assistant Professor in the Department of Computer Science, University of Sibiu, Romania.
      Dr. Linguraru’s research interests include medical image analysis, computer-aided diagnosis, multi-organ models of the human body, and translational methodologies for bench-to-bedside disease management. He is a member of the Technical Committee for Medical Imaging and Image Processing of IEEE Engineering in Medicine and Biology Society and recipient of a prize for Excellence in Engineering at the Houses of Parliament in London, UK. He has co-authored over 90 publications.
  • Homepage


問い合わせ先

  • 清水昭伸 (東京農工大学)
    • 電話 042 388 7478
    • 電子メール simiz @ cc.tuat.ac.jp

トップ   編集 差分 バックアップ 添付 複製 名前変更 リロード   新規 一覧 単語検索 最終更新   ヘルプ   最終更新のRSS
Last-modified: 2013-09-03 (Tue) 11:24:15 (1450d)