Classification of fluorescence microscopy
images of protein subcellular locations (PSLs).
The big picture goal of this project is to automate
the interpretation of fluorescence microscopy images. I was involved in
the classification step of the automation process. We aimed at
designing adaptive multiresolution features that are specially adapted
to the subcellular protien location patterns in a cell and hence
enhance the accuracy of automated classification.
Segmentation of multi-cell PSL images to obtain
single-cell images.
Once again as part of the process of automating the
interpretation of the images obtained by high-throughput fluorescence
microscopy, segementation, a pre-processing step for automated
classifcation, has to be automated. An algorithm based on the evolution
of stochastic active contours (STACS) was recently adapted to the
data set of 2D HeLa Images with great success. This has opened up
several venues of exploration and I will be joining the team on this
exciting adventure.
[Read more about these projects and related work on Prof. Jelena Kovacevic's web-page on Biomolecular Imaging.]