Events at SPIE Medical Imaging conference
SPIE 2010, SPIE 2009, and SPIE 2008SPIE Courses: Validation in Medical Image Processing (VMIP)
Instructor: Pierre Jannin, INSERM/Univ. de Rennes I (France)
This course provides attendees with a basic working knowledge of validation methodology in medical image processing. The course concentrates on explaining the main components of a validation study in medical image processing. Emphasis is given in designing and reporting validation studies. Practical and concrete advices are given for improving validation studies. Main examples concern image registration and segmentation. The second part of the course addresses the issue of assessment in a wider clinical application context. We outline the complexity and diversity of assessment in clinical contexts. Tools and classifications are introduced to manage this complexity. It is illustrated for Image Guided Interventions.
This course will enable you to distinguish main concepts in validation terminology and methodology, to design and report a validation study in medical image processing, to judge the quality of validation study results, and to identify the complexity of assessment of medical image processing in a targeted clinical area
This material is intended for anyone who needs to learn how to design and report a validation study of medical image processing and how to locate this validation in the global context of health care technology assessment. Those who either design their own medical image processing methods or who want to evaluate image processing in clinical context will find this course valuable.
SPIE 2007: WORKSHOP 'Validation in Image Registration Methods'Sunday 18 February, California Room · 5:45 to 7:45 pm
Chair: Pierre Jannin, INSERM/INRIA/Univ. de Rennes I (France)
Invited speakers: Daniel Rueckert, Imperial College London (United Kingdom); Gary E. Christensen, The Univ. of Iowa (USA) ; J. M. Fitzpatrick, Vanderbilt Univ. (USA)
- 5:45-6:15 : P. Jannin, INSERM/INRIA/Universite de Rennes (France): Methodology and terminology for reference-based validation of image registration Slides in PDF
- 6:15-6:45 : J. M. Fitzpatrick, Vanderbilt Univ. (USA) : Validating Rigid Registrations via fiducial markers Slides in PDF and in HTML: Rigid Registrations are registrations involving rigid-body transforms. These are the simplest transforms of interest in medical image registration. They are limited to images of rigid anatomy, such as bones, or the head, but because of their simplicity they are the easiest transforms to check against a standard. This talk will describe and discuss methods for determining the accuracy of rigid registration algorithms. It will focus on the method of comparing a transformation against a gold-standard transformation. It will suggest methods for measuring and reporting error and for estimating the effect of the standard.s errors on the measured error. It will include a case study of a project for validating rigid registrations between CT and MR images of the human head and will explore some of the difficulties faced in that project.
- 6:45-7:15 : D. Rueckert, Imperial College London (United Kingdom): Validation of non-rigid registration Slides in PDF: Over the last decade a large number of different algorithms for non-rigid registration have been developed. Even though these techniques have shown promising results, they are difficult to validate and compare. In this talk we will discuss different strategies for the task-specific validation and comparison of non-rigid registration techniques.
- 7:15-7:45 : G. E. Christensen, The Univ. of Iowa (USA) : Introduction to the Non-Rigid Image Registration Evaluation Project (NIREP) Slides in PDF: Non-rigid image registration (NIR) is an essential tool for morphologic comparisons in the presence of intra- and inter-individual anatomic variations. Many NIR methods have been developed, but are especially difficult to evaluate since point-wise inter-image correspondence is usually unknown, i.e., there is no "Gold Standard" to evaluate performance. The Non-rigid Image Registration Evaluation Project (NIREP) has been started to develop, establish, maintain, and endorse a standardized set of relevant benchmarks and metrics for performance evaluation of nonrigid image registration algorithms. This talk will describe the basic framework of the project.
The workshop was successful: 128 attendees until 8pm !!! Thanks for coming.
Acknowledgments: The invited speakers, Josien Pluim and Joseph Reinhardt for their invitation, Sandy for making handouts available during the workshop.