This is a personal homepage of Kijoon Lee, an assistant professor in bioengineering at Nanyang Technological University (NTU), Singapore. Nice to meet you in cyberspace.

- Last updated on November 22, 2007.

Kijoon who?

Kijoon is a south korean native, born and educated in Seoul. After receiving B.A. and M.S. in Physics from Seoul National University, continued research in nonlinear optics in Brown University, Rhode Island, U.S.A., where he obtained Ph.D in physics. Afterwards, he worked in the University of Pennsylvania as a postdoc, working on diffuse optical tomography for human breast cancer diagnosis. Currently he is an assistant professor in Division of Bioengineering, School of Chemical and Biomedical Engineering, in Nanyang Technological University, Singapore. For his curriculum vitae click here (pdf, doc)

His research interest is over broad area of biomedical imaging using diffuse light, including Diffuse Optical Tomography (DOT), Bioluminescence Tomography (BLT), and Diffuse Correlation Spectroscopy (DCS). These are emerging non-invasive imaging modalities that are capable of creating a 3D physiological map inside of a biological sample. Main motivation is to see a new physical model and instrumental ideas being translated into clinical environment, in order to raise the sensitivity/specificity of lesion detection and also raise patient comfort level.

Kijoon is currently teaching a tutorial on Biomedical Instrumentation for 3rd and 4th year students in bioengineering, SCBE, NTU.

Diffuse Optical Tomography (DOT)

Diffuse Optical Tomography, or DOT, is an emerging non-invasive clinical imaging modality that has vast amount of possibility. Light propagates in biological tissue in a diffusive manner due to high scattering from organells like mitocondria. When absorption is weak, one can model this light propagation using so called "photon diffusion equation". In near-infrared wavelength range (650-900nm) there exists a fortuitous spectral window in biological tissue where absorption is very small, so we can use this wavelength range to take multiple transmission measurement on different points on the sample surface. Since we have a simple model (photon diffusion equation) that predicts light intensity distribution for specific source positions under given optical property distribution (forward problem), we can attempt to reconstruct optical property distribution from the surface measurements (inverse problem). Since the inverse problem is ill-posed in general, there doesn't exist as well defined inversion algorithm as it does for X-ray CT or MRI. Image reconstruction in DOT is an art, which involves either linear or nonlinear optimization with a proper regularization. It is exciting as much as it is challenging, to do research in such a field where one is required to have knowledge in biology, physics, mathematics and computation at the same time.

Bio-Luminescence Tomography (BLT)

Unlike DOT where light source is provided externally from laser diodes, bioluminescence tomography, or BLT, makes use of internal light source, for example, luciferase expressed in a transgenic mouse. Map of bioluminescence can be obtained from surface imaging, but it only gives rough idea about which organ it is coming from. In order to get depth information as well, we need a model-based optimization for the source position and strength. Same photon diffusion equation can be used for this 3D reconstruction work, but this time the optical property distribution is assumed and we need to reconstruct the source strength at each voxel position.

Diffuse Correlation Spectroscopy (DCS)

DOT or BLT does not use coherent property of light - we only care about light intensity distribution. The coherence of light, though, can give us yet another information about the sample being probed. When we observe the intensity signal from a source-detector pair as a function of time, it shows different degree of random fluctuation depending on how much movement the inner scattering particles are undergoing. In other words, when a self-correlation curve is calculated, a faster moving sample will show faster decay than a slow moving sample does. It has been shown this correlation (as a function of source and detector position) satisfies tje good old photon diffusion equation. Therefore, one can attempt to obtain 3D map of flow (either Brownian or random) within the sample from surface measurements of intensity correlation. Application includes non-invasive functional imaging of brain and diagnostic imaging for peripheral arterial disease (PAD).