Archive for November, 2015

Seminar: Blind Compressed Sensing in Biomedical Imaging


Dr. Yoram Bresler, professor at the University of Illinois at Urbana-Champaign, presented his work "Blind Compressed Sensing in Biomedical Imaging: data-adaptive sparse modeling and acquisition"

According to Dr. Bresler, compressed sensing methods have generated tremendous interest in biomedical imaging: they can produce high quality images from significantly reduced data acquisition. He showed that replacing fixed models of sparsity and randomized acquisition in these methods by data-driven learning can help unleash the full potential of compressed sensing. These approaches are illustrated on CT and MRI.

Dr. Yoram Bresler, Professor
Coordinated Science Laboratory and the Department of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign

Professor Bresler delivering his presentation

Professor Bresler delivering his presentation

Seminar: Medical Image Synthesis


Dr. Jerry L. Prince presented his work: "Medical Image Synthesis: Methods and Applications".

According to Dr. Prince, the acquisition of truly calibrated magnetic resonance images is not currently possible. Scanners and pulse sequences are different—in subtle ways sometimes and quite dramatic ways more often. Manufacturers have different strategies for optimizing their image quality and MR techs might change a parameter to try to improve the image quality on any given day. As a consequence, the use of MR images for automatic image analysis yields inconsistent results. We have been exploring image synthesis methods to address this problem and are hopeful that through synthesis we will be able to obtain more consistent image analysis results. If successful, the use of automatic image analysis methods applied to MR images might become a more important part of clinical practice in the future. Four methods and a variety of applications and their results are presented in this talk. Sparse reconstruction is an important theme throughout, and image segmentation and registration are key methods that serve to demonstrate improvements. Although our results are promising, this new area of research is controversial and its future impact is uncertain. Dr. Prince concluded with some ideas about future directions and some thoughts about what might be possible in the future.

Dr. Jerry L. Prince
William B. Kouwenhoven Professor
Electrical and Computer Engineering
Johns Hopkins University

Dr. Prince delivering his presentation at the Gordon Center

Dr. Prince delivering his presentation at the Gordon Center

IEEE-NPSS/MIC Conference Awards 2015

Two Gordon Center members were honored at the IEEE Nuclear and Plasma Sciences Society / Medical Imaging Conference.

Early Achievement Award

Quanzheng Li, Ph.D., received the IEEE-NPSS Early Achievement Award for outstanding achievement to the field of medical imaging science, "for contributions to the development of a MAP framework for iterative image reconstruction on static, dynamic, parametric and TOF PET, and its application in whole-body dynamic imaging and computer-aided detection."

The 2015 Bruce Hasegawa Young Investigator Medical Imaging Science Award

Se Young Chun, Ph.D., received the 2015 Bruce Hasegawa Young Investigator Medical Imaging Science Award from IEEE Nuclear and Plasma Sciences Society (NPSS) at the IEEE Medical Imaging Conference, San Diego, CA, USA. He has been selected as this year’s recipient of this award for contributions to image reconstruction methods in the presence of object motion.

This annual award is given to young investigators to recognize their contributions to the medical imaging field.

Dr. Se Young Chun receiving his award

Se Young Chun receiving his award

IEEE-MIC Oral: PET/MRI Joint Estimation

A Joint Estimation Method for Kinetic Modeling of Simultaneously Acquired PET/MRI Signals

Moses Q. Wilks, Xiaomeng Zhang, Jinsong Ouyang, Georges El Fakhri, Nathaniel M. Alpert, Quanzheng Li.

At the IEEE Medical Imaging Conference, Dr. Wilks presented a method to jointly estimate kinetic parameters from simultaneously acquired PET and MRI data, using the Alternating Direction Method of Multipliers (ADMM). Often, data acquired on the relatively new modality of PET/MRI scanners is used independently. This method was developed to use this simultaneously acquired data in a synergistic fashion. Computer simulations showed that it could improve model fitting in the case of CTE-MRI and 15O-H2O PET data. This was extended to in vivo imaging of a rabbit implanted with a tumor, and simultaneously imaged with Magnevist as a MRI contrast agent and 15O-H2O. A 4x4x5cm region of interest was created over the tumor and vascular extraction (K1) rates were measured using standard techniques (PET Alone), or by simultaneous modeling (ADMM). Due to the noisy nature of the 15O-H2O scan, PET alone was not sufficient to create reasonable measurements, but using ADMM structural and biological information was retrieved.

Parametric maps of K1 using PET alone or PET with MRI (ADMM method)

Parametric maps of K1 using PET alone or PET with MRI (ADMM method)