Research Projects

Amygdala/Hippocampus Modeling Pipeline
April 16, 2008

We have developed a unified image processing and analysis pipeline for brain substructures for amygdala (Chung et al., 2008, MIAR) and extending the pipeline to hippocampus. This is a joint collaboration with Richard J. Davidson and Brendon M. Nacewicz of the Waisman Laboratory for Brain Imaging and Behavior, and with Seth Pollak and Jamie Hanson of the Child Emotion Research Laboratory.

Basis Method for Arbitary Manifolds
April 16, 2008-

This is a joint project with Anqi Qiu of the National University of Singapore. For parametric model building on arbitary manifolds, it is necessary to construct an orthonormal basis. Previously eigenfunctions of the Laplace-Beltrami operator has been used as a natural choice (Qiu et al. 2006, TMI). We have developed a  much simpler technique that avoids solving a large system of equations.

Weighted Spherical Harmonic Representation
2006-

The traditional spherical harmonic representation suffers from the Gibbs phenomenon. By simply adding exponential weights to the representation, we can make the representation converges faster and reduce the Gibbs phenomenon. We have collaborated with Li Shen of the Indiana University School of Medicine, on developing an iterative algorithm for constructing the representation for large scale anatomical structures, as well as Richard Hartley of Australian National University. The project resulted in at least four major publications (Chung et al. 2006. MMBIA; 2007. TMI; 2008. TMI; 2008. Statistica Sinica).

Diffusion Tensor Imaging
2002-

With JeeEun Lee and Andrew L. Alexander of the Waisman Laboratory for Brain Imaging and Behavior, we have been developing various DTI anisotropic filtering technique. Based on the idea that anisotropic kernel smoothing can approximate the solution of the isotropic diffusion equation (Lee et al., 2006. ISBI), we smooth out DTI.

Brain and Behavior Correlation in Autism
2002-

We have been collaborating with  Richard J. Davidson, Kim M. Dalton, Daniell J. Kelley of the Waisman Laboratory for Brain Imaging and Behavior. Both the brain functional and structural images are correlated with various behavioral measures in quantifying deficits in autism (Chung et al., 2004. NeuroImage; 2005. IPMI; 2005, NeuroImage).

Tensor-based Morphometry
2001-

Using the diffeomorphic registration (Avants et al., 2006), we are trying to quantify abnormal growth in instutalized children. The joint callobration with James Gee and Brian Avant of the University of Pennsylvania, and Seth Pollak and Jamie Hanson of the Child Emotion Research Laboratory. We have currently finished preprocessing approximately 70 images.

Longitudinal Vocal Tract Structure Modeling
2006-

Using the combined CT/MRI images of more than 300 subjects, we longitudinally model the growth of various vocal tract structures in a multivarite fashion. Joint project with Houri K. Vorperian of the Vocal Tract Development Laboratory. This collaboration resulted in three journal papers so far.

Mandible Shape Analysis
2007-

Using about 300 CT images of mandible, we are trying to build a mandible shape model on binary segmentation. Joint project with Houri K. Vorperian of the Vocal Tract Development Laboratory. Currently an initial image processing has been finished. We have applied NIH grant for a further analysis.

Rosenthal Fiber Segmentation
Febuary 28, 2008-

Using a modified Gaussian mixture model with the expectation maximization algorithm, the Rosenthal fibers in rat confocal microscopic images can be segmented. This is a joint project with Jolien Conner and Albee Messing of the Waisman Center.