There’s a myriad of software tools that can detect lines, shapes, and edges within medical scans. Most of the effort within the field of image segmentation has been focused on developing fully automatic algorithms, and overall they’ve gotten impressively good. Nevertheless, the human body keeps challenging the available software, requiring a lot of radiologists’ time to confirm and analyze what the software outlines. Researchers at Ben-Gurion University of the Negev in Israel developed segmentation software that partners with the user to produce better results.
While some software may decide that a region is not part of a tumor, using the new application an astute doctor that knows better would place a dot on that area. This tells the application to include that region within the nearby segment and an “influence” slider tells the software how strongly the user feels about that decision. It then takes those parameters into account and runs its segmentation algorithms accordingly.
Check out this silent video that demonstrates how the software works:
The research was conducted by Dr. Ellan Shelef, Head of Diagnostic Imaging Institute at Soroka Medical Center, Ben-Gurion University of the Negev, and Dr. Tammy Riklin-Raviv, Head of BioMedical Image Computing (BioMic) laboratory, Electrical and Computer Engineering Department of Ben Gurion University of the Negev (rrtammy@bgu.ac.il)
Mr. Itzik Mashiach handles Ben Gurion Negev Technologies Transfer program (itz@ovelatech.com)