After completing this course, students will be able to:
- understand the importance and role of quantitative image analysis in clinical research and practice. This will include getting familiar and proficient with open source medical image exploration tools.
- recognize clinical issues and problems that can be solved effectively and efficiently with a machine learning approach. They will also be able to identify and understand the utility of machine learning in diagnosis and in implementation of personalized medicine, and to fluently and effectively select appropriate tools and interpret their outcome
- understand the fundamentals of deep learning, its relevance to machine learning, and how it can be used in lieu of other open source tools (above) for image processing, registration, and interpretation
- specify functional and structural architecture and basic design to implements deep learning models. They will be able to implement and/or extend appropriate deep learning architectures using Python
- identify, understand, and quantify the limitations of deep learning architectures, implementations, and application
- recognize clinical issues that can be solved effectively and efficiently with a 3D technology approach, and acquire, create, and manipulate 3D mesh data.
Broadly, after attending this course students will be able to assess whether a clinical problem can benefit from machine learning and/or deep learning or from 3D technologies. Students will be able to implement machine learning and deep learning approaches to scientific research, clinical practice, clinical diagnosis, and individualized medicine. They will also be able to administer and/or utilize these technologies in their clinical practice and research.|
The emphasis of this course is on understanding of the principles, implementation, and limitations related to these technologies, and on developing insights about their use to address clinical problems. A full, rigorous engineering or mathematical treatment of these technologies is beyond the scope of this course.
This course includes several assignments that will be completed as randomly assigned pairs of students, and will including several sub-assignments as appropriate. Sub-assignments will not be graded individually but will be used to provide feedback to the students and to assess teaching points that require attention. The course will also include a 3D visualization project that covers 3D image acquisition, processing, and interpretation in clinical setting. This course does not include a written final exam as the focus in on practical aspects. Knowledge base will be assessed as a component of the assignments.
This course is open only to TM students.
Visiting professor (Massachusetts General Hospital and Harvard Med School): Gupta, R..
Gastdocent (MST): Doremalen, Rob van
Gastdocent (RUMCN): Maal, T.J.J.
Assumed previous knowledge
Segmentation and Visualization (201200168-2A)
|Master Technical Medicine||Required materials|
Recommended materials-Instructional modes
|Study materials will include lecture notes, handouts distributed as part of the assignments, links to or copies of instructions for necessary software installation and use, and lists of readings (chapters from publicly available online books) for each lecture as applicable and appropriate.|
|Work books: Pdf forms will be used for exercises|
|Python environment (in the form of access to UT JupyterLab -- students should already have access, but will be covered in one of the early courses if necessary).|
|Self study with assistance|
Remark4 assignments (20 points each, minimum 10
1 project (20 points, minimum 10)