Ohio State Navbar

The Ohio State UniversityOffice of International Affairs

give a donation

Quantification of Biological Structures from Images of Immunostained Slides

Research Scholar

M Khalid Khan Niazi, Biomedical Informatics (Pakistan)
Anjali Satoskar, Co-Researcher
Metin Gurcan, Faculty Mentor


M Khalid Khan Niazi is a postdoctoral researcher with Metin Gurcan in the Clinical Image Analysis Laboratory at The Ohio State University. He got his PhD in medical image processing from Center for Image Analysis at Uppsala University, Sweden in 2011. His research interests include directional analysis, multi-scale analysis, image registration, stationary/non-stationary image analysis and shape analysis. Currently, his research is focused on quantification of different biological structures in immunohistochemical stained images.

What is the issue or problem addressed in your research?

The quantification of different biological structures in immunohistochemical stained images plays a vital role in digital pathology. The quantification is often performed visually, a process which is tedious and subject to inter- and intra-reader variability. To standardize this practice, we are developing image analysis methods that can efficiently perform quantification of different biological structures from the scanned images of immunostained slides. For instance, counting the number of cytotoxic t-cells from scanned images of allograft biopsy is often considered as an important clue to differentiate between renal infection and rejection. However, counting cytotoxic t-cells in presence of cell clumps, irregular shape of the stained regions, color variation, huge size and shape variation of the cells, becomes a challenging task for both the pathologist and the computer. To overcome these issues, we have developed an automated method that assists the pathologist by counting the number of cytotoxic t-cells from stained images of allograft biopsy.

What methodology did you use in your research?

The method exploits the intrinsic properties of the color space to segment the cytotoxic t-cells from the stained images. Once segmented, the cytotoxic t-cells are counted by the fusion of stable points extracted from different color channels. Stable points are extracted by using the normalized Laplacian of Gaussian scale space framework. The stable points between the channels are fused based on the strength of the curvature information of these points.

What are the purpose/rationale and implications of your research?

The main purpose of this research is to develop image analysis methods which will assist the biologists/clinicians/pathologists in making accurate assessment, diagnosis, and conclusions about different medical conditions. It will help in eliminating the user bias and will reinforce standardization of analysis among research labs.