This is the title of a presentation by Michelle Borkin (Harvard University), given on February 5, 2013, as part of STScI’s Engineering and Technology Colloquium series.
Michelle’s specialty is in developing new approaches to interdisciplinary scientific visualization and data exploration. She has been developing new visualization techniques in medical imaging, especially in the field of cardiology, and is a member of the “Astronomical Medicine” project at Harvard. Her talk at STScI concerned was concerned with the connection between medical and astronomical imaging.
The first part of the talk described work in developing new techniques to visualize the structure of coronary arteries to determine whether plaque is building up on the walls. 3-D plots of the heart, derived from CAT scans, are visually impressive, but the arteries can obscure each other and make diagnosis harder. Michelle developed a tree-leaf visualization technique – shown below – that revealed the full details of the structure of the article, and trials indicated fewer misdiagnoses with this new approach.
Here is a 3-D image:
… and here is the tree-leaf representation:
The work showed the importance of consulting with customers. Examples: the organization of the tree-leaf structure reflected the way doctors think about the heart and how it operates, and the color table reflected how they think about data: a gray-scale (while showing more detail) was assumed to show X-ray data.
The same techniques can be applied to hierarchical astronomical data – gravitational contraction of a molecular cloud, leading to star formation. As a result of this work, new tools are being developed for exploring astronomical data. One of these is Glue, a Python library aimed at linking and exploring data, and it supports Python scripting as well. Listen to Chris Beaumont explain how it works:<p><a href=”http://vimeo.com/53378575″>What is Glue?</a> from <a href=”http://vimeo.com/beaumont”>Chris Beaumont</a> on <a href=”http://vimeo.com”>Vimeo</a>.</p>