RESEARCH EXPERTISE

Image segmentation and registration, deformable models, image analysis, digital topology, mesh simplification, super-resolution reconstruction, variational and PDE methods and medical imaging.

EDUCATION

Johns Hopkins University, Baltimore, MD

Ph.D., Electrical and Computer Engineering, (GPA: 3.95/4) expected February 2008

Dissertation title: Accurate, Efficient, and Topologically Correct Cerebral Cortex

Reconstruction from Magnetic Resonance Images

Advisor: Professor Jerry L. Prince

M.S.E., Electrical and Computer Engineering, May 2005

Shanghai Jiao Tong University, Shanghai, China

B.S., Electrical Engineering June 2002

EXPERIENCE

Siemens Corporate Research, Princeton, NJ

Intern Imaging and Visualization Department May 2005-Aug 2005

Developed a nearly automatic segmentation method for extraction of esophagus from CT images using a probabilistic shortest path approach that integrates prior and contextual information.

Johns Hopkins University, Baltimore, MD

Research Assistant Electrical and Computer Engineering Department Jan 2003-present

Designed and implemented a MAP super-resolution method to reconstruct a high-resolution MR image from two orthogonal scans of the same subject.
Proposed and validated a novel fuzzy segmentation framework that incorporates the idea of super-resolution image reconstruction.
Designed and implemented a PDE diffusion based method for zipper artifact correction in MR brain images.
Established a new digital topology theory on adaptive octree grids. Constructed rigorous proofs. Designed practical implementation algorithms.
Designed an octree-based isosurface generation and simplification method that preserves topology, guarantees no self-intersections, and generates a multi-resolution surface that approximates the true isosurface of the underlying data.
Developed a topology-preserving geometric deformable model on adaptive octree grids for 3D medical image segmentation. Applied the model in cortical surface reconstruction with improvement in both accuracy and efficiency.

Johns Hopkins University, Baltimore, MD

Teaching Assistant Electrical and Computer Engineering Department Sept 2005-Dec 2006

Medical Image Analysis: Designed final exam projects, Graded homework, held office hours, and led review sessions.
Random Signal Analysis: Graded homework, held office hours, and led review sessions.

SKILLS

Programming: C/C++, MATLAB, UNIX Shell Scripts, perl, HTML.

Computer Skills: Unix/Linux, Windows, Office, Visual Studio.

Languages: English (fluent), Chinese (native).

PROFESSIONAL ACTIVITIES

Journal Reviewer:

· IEEE Transactions on Image Processing

· IEEE Transactions on Medical Imaging

· Visual Computer

· Medical Physics

Organization Affiliation:

· IEEE Student Member

PUBLICATIONS

Submitted journal articles:

Y. Bai, X. Han, and J. L. Prince, "Digital Topology on Adaptive Octree Grids," (submitted to Journal of Mathematical Imaging and Vision).

Journal articles in preparation:

Y. Bai, X. Han, and J. L. Prince, "Adaptive Resolution Topology Preserving Geometric Deformable Model for Cortical Segmentation".
Y. Bai, X. Han, and J. L. Prince, "Super-resolution Tissue Classification of MR Brain Images from Orthogonal Low Resolution Scans".

Book chapters:

Y. Bai, X. Han, and J. L. Prince, "Geometric deformable models: overview and recent developments," Biomedical Image Analysis: Methodologies and Applications (invited).

Conference papers:

Y. Bai, X. Han, and J. L. Prince, "Octree Grid Topology Preserving Geometric Deformable Models for 3D Medical Image Segmentation," International Conference on Information Processing in Medical Imaging (IPMI'07), July 2007.
Y. Bai, X. Han, and J. L. Prince, "Topology-preserving Geometric Deformable Model on Adaptive Quadtree Grid," IEEE Proc. Conf. on Comp. Vis. Patt. Recog. (CVPR'07), June 2007.
Y. Bai, X. Han, and J. L. Prince, "Octree-based Topology-preserving Isosurface Simplification," IEEE Workshop on Mathematical Methods and Biomedical Image Analysis (MMBIA'06), June 2006.
M. Rousson, Y. Bai, C. Xu, and F. Sauer, “Probabilistic Minimal Path for Automated Esophagus Segmentation,” Proc. SPIE Medical Imaging, pp. 614449-61444H, 2006.
Y. Bai, X. Han, and J. L. Prince, "Super-resolved Multi-channel Fuzzy Segmentation of MR Brain Images", Proc. of SPIE Medical Imaging, 2005.
Y. Bai, X. Han, and J. L. Prince, "Super-resolution Reconstruction of MR Brain Images," Proc. of 38th Annual Conference on Information Sciences and Systems (CISS'04), March 2004.

Dissertation:

Y. Bai, "Accurate, Efficient, and Topologically Correct Cerebral Cortex Reconstruction from Magnetic Resonance Images," Johns Hopkins University, Baltimore, MD, 21218, USA, in preparation.

Patent:

"Probabilistic Minimal Path for Automated Esophagus Segmentation", M. Rousson, C. Xu, and Y. Bai, US Pending 11/481,992.