Program
8:00 – 8:30 am: Registration, speaker check-in and poster setup
8:30 – 8:45 am: Opening Remarks
8:45 – 9:30 am: Keynote Speaker:
Dr. Jerry Prince, Professor, Electrical and Computer Engineering, Johns Hopkins University
Title - "Patches and pulse sequences for MR image intensity normalization"
Abstract - "Despite years of effort and many proposed methods, magnetic resonance images continue to lack a standard intensity scale. As a result,algorithms designed for one pulse sequence and scanner strength will typically fail to perform as expected when images acquired with (even slight) variations from the original design specifications are used as input. While tweaking parameters or retraining can improve results,it may be more practical in some cases to add a preprocessing step to adjust the intensity scale of the acquired images to resemble that of the design specs--this is the process of intensity normalization. An approach called PSICLONE, which uses patches for image synthesis, is described. The twist here is that while an atlas is presumed to be available, an additional image that is synthesized using estimated pulse sequence parameters is added to the atlas. Patches in this image together with the remaining images in the atlas are used to train a random forest that can synthesize any other tissue contrast in the atlas. After describing the approach in detail, various validations, examples, applications, and potential future directions will be provided."
9:45 – 10:30 am: Keynote Speaker:
Dr. Gary Christensen, Professor, Electrical and Computer Engineering, University of Iowa
Title - "Current- and Varifold-Based Lung Image Registration"
Abstract - "Registering lung CT images is an important problem for many applications including radiation therapy planning and treatment, tracking lung motion over the breathing cycle, tracking anatomical and function changes over time, and detecting abnormal mechanical properties of the lung. This talk will explain how currents and varifolds can be used to represent tree-like structures and surfaces of the lung. In this approach, curve-like structures in the lung---for example, the skeletons of vessel and airway segmentations---are represented by currents or varifolds in the dual space of a Reproducing Kernel Hilbert Space (RKHS). In a similar manner, we will show how surfaces of the lung can be represented via currents or varifolds. Current and varifold representations are implemented using discrete Dirac delta currents/varifolds that are parameterized via of a collection of momenta. This discrete current/varifold representation can be viewed as a patch-based representation for image registration. Next, we will discuss how discrete current/varifold representations can be registered using the Large Deformation Diffeomorphic Metric Mapping (LDDMM) framework. This framework produces correspondence maps (transformations) between images that are guaranteed to be diffeomorphisms. In this approach, the velocity field of the LDDMM framework is represented via currents parameterized by momenta. Finally, we will compare and contrast current- and varifold-based diffeomorphic image registration approaches using simple 2D phantom images and real 3D CT lung images."
10:30 – 11:00 am: Coffee break and poster session
Title | Authors |
---|---|
AAutomatic Segmentation of Hippocampus for Longitudinal Infant Brain MR Image Sequence by Spatial-Temporal Hypergraph Learning | Yanrong Guo, Pei Dong, Shijie Hao, Li Wang, Guorong Wu, and Dinggang Shen |
Construction of Neonatal Diffusion Atlases via Spatio-Angular Consistency | Behrouz Saghafi, Geng Chen, Xi’an, Feng Shi, Pew-Thian Yap, and Dinggang Shen |
Selective Labeling: identifying representative sub-volumes for interactive segmentation | Imanol Luengo, Mark Basham, and Andrew French |
Consistent multi-atlas hippocampus segmentation for longitudinal MR brain images with temporal sparse representation | Lin Wang, Yanrong Guo, Xiaohuan Cao, Guorong Wu, and Dinggang Shen |
Sparse-Based Morphometry: Principle and Application to Alzheimer’s Disease | Pierrick Coupé, Charles-Alban Deledalle, Charles Dossal, and Michele Allard |
Multi-Atlas Based Segmentation of Brainstem Nuclei from MR Images by Deep Hyper-Graph Learning | Pei Dong, Yanrong Guo, Yue Gao, Peipeng Liang, Yonghong Shi, Qian Wang, Dinggang Shen, and Guorong Wu |
Non-local MRI Library-based Super-resolution: Application to Hippocampus Subfield Segmentation | Jose Enrique Gómez, Pierrick Coupe, and Jose Manjon |
Patch-based DTI grading: Application to Alzheimer's disease classification | Kilian Hett, Vinh-Thong Ta, Rémi Giraud, Mary Mondino, Jose Manjon, and Pierrick Coupé |
HIST: HyperIntensity Segmentation Tool | Jose Manjon, Pierrick Coupe, Parnesh Raniga, Ying Xia, Jurgen Fripp, and Olivier Salvado |
CapAIBL: Automated reporting of cortical PET quantification without need of MRI on brain surface using a patch-based method | Vincent Dore, Pierrick Bourgeat Victor Villemagne, Jurgen Fripp, David Ames, Lance Macaulay, Colin Masters, Christopher Rowe, and Olivier Salvado |
High resolution hippocampus subfield segmentation using multispectral multi-atlas patch-based label fusion | Jose Enrique Gómez, Pierrick Coupe, and Jose Manjon |
Identification of water and fat images in Dixon MRI using aggregated patch-based convolutional neural networks | Liang Zhao, Yiqiang Zhan, Dominik Nickel, Matthias Fenchel, Berthold Kiefer, and Sean Zhou |
Estimating Lung Respiratory Motion Using Combined Global and Local Statistical Models | Zhong Xue, Ramiro Pino, and Bin Teh |
11:00 am – 12:45 pm: Oral Presentations:
Title | Authors |
---|---|
Robust and Accurate Appearance Models based on Joint Dictionary Learning: Data from the Osteoarthritis Initiative | Anirban Mukhopadhyay, Oscar Morillo, Stefan Zachow, and Hans Lamecker |
Patch-Based Discrete Registration of Clinical Brain Images | Adrian Dalca, Andreea Bobu, Natalia Rost, and Polina Golland |
Hierarchical Multi-Atlas Segmentation using Label-Specific Embeddings, Target-Specific Templates and Patch Refinement | Christoph Arthofer, Paul Morgan, and Alain Pitiot |
Supervoxel-Based Hierarchical Markov Random Field Framework for Multi-Atlas Segmentation | Ning Yu, Hongzhi Wang, and Paul Yushkevich |
12:45 – 1:00 pm: Closing remarks and best paper award