My most recent research interests are:
Tree and graph structures consisting of tubular segments are omnipresent in biological systems; consider the axons and dendritic trees of neurons, vascular networks in various organs, bronchial tree in the lungs, river and road networks in geospatial imagery (a 2-dimensional instantiation of the same concept). Biomedical 3-dimensional (3D) imagery obtained using various imaging modalities including confocal microscopy, computerized tomography (CT), magnetic resonance imaging (MRI) can be utilized to analyze and extract the 3D structure of tubular object networks which could be used to model and simulate such networks structurally and functionally, as well as design virtual endoscopy type diagnostic applications. We are currently focusing on confocal microscopy imagery of neural networks and their dendritic tree structures. Especially, we have been working with Brainbow imagery. Automatic extraction of dendritic tree structures is an important first step in processing massive image data to be used in improving our understanding of how the brain and the nervous system works. Our submission to the DIADEM challenge is advanced to the final round! You can find the details here. For more info ->.
Analysis of lipid vesicles
Collobrated work with the researches at University of Ljubljana. Details are in JBO'11 paper. A nice video prepared and presented by Jernej:
Image-guided tumor tracking for radiation therapy
We are working on data fusion of KV and CT images to build a robust real time tumor tracking model. In order to increase the accuracy of radiation therapy, our method utilizes the accurate geometry information from the CT scans to build a motion model, which is then used for real time tracking on KV images that are acquired on the treatment day. To achieve this, we first register KV and CT images on a common coordinate system. For that purpose, reliable and fast Digitally Reconstructed Radiographs (DRRs) are acquired from CT images, using edge preserving ray casting. Then, in order to register 2D KV images with 2D DRRs, a nonparametric kernel density estimation (KDE) based approach, that performs well at low SNR and is robust to the missing data, etc., can be used.
My previous research interests:
Indoors localization, activity classification, and behavioral modeling are increasingly important for surveillance applications including independent living and remote health monitoring. For such applications, fish-eye cameras are very promising for accurate assessment of complex behavior and activity patterns with the proper application of video processing and computer vision techniques. I implemented a multi (2) camera system that can be used for tracking of multiple people under heavy occlusion, luminance change, human-human and human-object interaction.
Image segmentation and analysis of MR sequences
To investigate spatial distribution of iron accumulation in globus pallidus (GP) in patients with Hallevorden-Spatz Syndrome (HSS) also known as pantothenate kinase-associated neurodegeneration (PKAN), phase imaging method can be used. Our results demonstrate higher sensivity of the phase measurements for quantitative assessment of iron concentration compared to the relaxation rate measurements (R1, R2, and R2*).
I worked on reconstruction methods for T1 mapping in dGEMRIC technique. By restoring the signal polarity in IR sequence, it is possible to use the whole dynamic range of the data and improve the reliability of T1 relaxation time fits to an inversion-recovery function.
Traffic Analysis from video (MSc Thesis)
My MSc Thesis was titled Road and Traffic Analysis from Video. I developed 2 video-based traffic analysis systems: one for traffic monitoring with fixed cameras, and one for driver warning applications with on-board cameras looking outwards from the windshield. Here are sample videos.
and lastly, here is my wordle