Surya Prasath, PhD, an expert in imaging informatics, image processing and computer vision, has joined the biomedical informatics team at Cincinnati Children’s and the University of Cincinnati as an Assistant Professor. Previously with the University of Missouri-Columbia Department of Computer Science, he holds a PhD in Mathematics from the Indian Institute of Technology in Madras, India.
Prasath has already been collaborating with Cincinnati Children’s Bruce Aronow, PhD, and his team in conducting large-scale histopathological image analysis of brain tumors utilizing datasets from the National Cancer Institute and Cincinnati Children’s. Here, he tells us a little more about his research goals:
What drew you to want to work at Cincinnati Children’s?
I wanted to join an environment that fosters autonomy, passion and creativity. Cincinnati Children’s, in my opinion, is at the top in terms of giving early career researchers a first-class environment to develop essential skills in advancing medical and clinical care research. Also, they offer the unique advantage of a highly multidisciplinary network of collaborators, including computer scientists, neuroscientists, radiologists, pathologists and physicians.
Could you describe your research?
My main lines of research are in mathematical analysis and image processing for various biomedical imagery modalities. I have worked with imagery ranging from micro-microscopy and histopathology to macro-MRI and endoscopy. Biomedical image informatics research is crucial in medical research and clinical care, as automatic image analysis tools can provide valuable insight into various diseases and help in diagnostics.
Some of my current translational biomedical imagery based projects include:
- Computational pathology: State of the art segmentation and classification methods for breast cancer and glioma brain tumor histopathological tissue images
- Microscopy image analysis: Filtering, denoising, segmentation and classification methods for fluorescence/confocal/photon microscopy with deep learning techniques
- Radiology imaging informatics: Automatic algorithms for segmentation of neonatal brain tissues and measurement of corresponding volumes using T1- and T2-weighted MRI scans
What are your research goals?
My goals are to develop automatic biomedical image analysis tools, emphasizing the integration of imaging informatics data with clinical, genetic and genomic information for precision medicine and research. One focus will be on designing and developing software for processing large-scale spatiotemporal biomedical data sets in an effective and clinically meaningful way. Further, I would like to create and enhance data-mining tools with integrated image analysis modules and high-performance computations (HPC).
How did you become interested in imaging analysis?
Visual data excites me, and I am slowly turning into a data scientist now. I started in pure mathematics, focusing during my master’s degree and the first few years of my PhD on learning the core topics in mathematics such as analysis, topology and geometry. Now in the past decade, I’ve spent a lot of time in learning image processing and computer vision techniques related to biomedical image analysis. I bring in a strong mathematical background that enables me to delve deeper into automatic image analysis techniques.
What do you want other researchers at Cincinnati Children’s to know about you?
I am always interested in making connections in different areas, especially given the proliferation of biomedical visual data. I would like to connect with other researchers at Cincinnati Children’s to explore and extract meaningful information from those massive datasets.
What excites you most about the future of biomedical informatics?
We are living in the golden age of unprecedented technological innovation. In computer science, artificial intelligence and machine learning research have provided exciting developments in recent years. The ramifications of machine learning are enormous, especially with the current wave of deep learning techniques to extract salient information from big data. We are already seeing systems developed than can outperform the domain experts in computational pathology and radiology informatics. I am particularly excited about the future of biomedical image and video analysis, where we now have both computational resources and cutting-edge machine learning techniques to obtain meaningful results.