Research Areas and Topics
Current projects in the field of graphics and real-time rendering include markerless 3D motion estimation from videos, artistic style characterization for non-photorealistic rendering applications, visually realistic simulation of complex underwater environments, accurate 3D mesh reconstructions of lung/brain anatomy from HRCT/MRI image stacks, and applications of dual-quaternions in skeletal animation.
The research group is currently involved in several collaborative projects with clinicians and radiologists. Key research areas are analysis of whole slide images (WSI) of breast cancer histology slides, mRNA tissue micro-array analysis using BRCA1, BRCA2 biomarkers, ultrasound image analysis algorithms for speckle reduction, feature enhancement and classifcation, digital mammogram image analysis using multifractal descriptors, and MRI image analysis and segmentation algorithms for Parkinson's disease characterization.
A very important problem in the field of digital pathology is the automated classifcation of images for computer aided diagnosis of diseases. We are currently working on several machine learning algorithms for HER2 classification in breast cancer histology slides using biomarker specific features. Other algorithms being studied are emphysema classification in HRCT images, and characterizations of microcalcifications in digital mammogram images.