By Christian Stafford
Third-year CCIS PhD student Aditeya Pandey was recently named the recipient of the 2018 Institute of Electrical and Electronics Engineers Visualization Conference (IEEE VIS) Best Poster award for his research on the design of a visual Clinical Decision Support Systems (CDSS) – a suite of software that helps healthcare providers make better decisions and improve patient care.
IEEE is the world’s largest association of technical professionals with more than 423,000 members in over 160 countries around the world. The association’s VIS conference serves as the most important conference on scientific visualization, information visualization, and visual analytics – it is the premier forum among visualization practitioners in academia, science, government and industry.
Pandey’s research poster, titled, “CerebroVis: Topology and Constraint-based Network Layout for the Visualization of Cerebrovascular Arteries,” illustrates the significance of the CerebroVis tool that Pandey and his team created – “a novel, 2D network-visualization layout that represents the cerebral network, designed to assist doctors in the diagnosis of cerebrovascular diseases,” Pandey said. “The layout design enables the doctor’s routine diagnostic-like examination of artery branches for abnormalities and the identification of paths of abnormal flow from a deformed artery.”
The research into CerebroVis is significant in that it can assist doctors in the diagnosis of blood flow related diseases such as strokes and aneurysms. Pandey explained, “Our collaborators like the tool because the tool is designed as an “assistant” and helps them in their task rather than telling them what should be done.”
Pandey’s advisor, CCIS Assistant Professor Michelle Borkin, Ph.D., has a long-time interest in astronomy and physics, and she worked on developing visualization tools for NASA, where she eventually “fell in love with computer science and data visualization,” she said. Commenting on Pandey’s achievement, Borkin emphasized that “The purpose of the research is to help doctors visualize arteries in the human brain, and present artery structures so doctors can more accurately and efficiently diagnose diseases in the brain such as blood clots, strokes and hemorrhages.” Simply put, the CerebroVis tool helps doctors identify brain diseases “by looking at changes in the patterns of blood-flow in the arteries,” said Borkin.
Cody Dunne, assistant professor in CCIS, who spent three years working at IBM in their research, Watson, and Watson Health divisions, also supervised Pandey’s research. “We have a lot of evidence demonstrating that 3D network visualizations are less effective than 2D representations for topology and attribute-based user tasks.” He explained, “There are challenges however in flattening a 3D structure like brain arteries in to a 2D space. Think, for example, of the distortions caused by flattening a globe into a map. Placing CerebroVis into a larger context, Dunne added, “just as different map projections are better or worse for different user tasks, the 2D layout we choose for arteries will have an impact on how efficient and accurate our users will be.’
In addition to advancing the research he has already conducted, one of Pandey’s goals is to look further into the development of human-centered visualization systems for the medical science community. “I would like to invest time into developing my research prototype into an application that will have a greater outreach in the community,” he said.
With an eye on the future, Pandey said he would like to work on projects that have a social impact and increase the boundaries of technology. “CCIS is growing its data science research ecosystem and I look forward to interesting research collaborations coming out of this environment.”