Paper Reading #23: Evaluating Automatic Warning Cues for Visual Search in Vascular Images

References:
Title: Evaluating Automatic Warning Cues for Visual Search in Vascular Images
Authors: Boris van Schooten, Betsy van Dijk, Anton Nijholt, and Johan Reiber
Venue: IUI 2010, Feb. 7-10 2010

Summary:
This paper was about automatic warning systems being implemented with visual search systems. The visual search system used in this study is a vascular structure. The application used in this search involving vascular image analysis is a 3D magnetic resonance angiography (MRA) which is used by radiologists.

MRA of the brain and part of the spine

The tests were done with systems that provided four different conditions. The system either produced no warnings, only false positives, only false negatives, or a perfect system that had no errors at all. It was shown that the idea of more false positives was best and preferred.  It was also shown that the users did not mind more warnings on the screen.

The picture comes from cedairs-sinai.edu

Discussion:
This idea is really important because can help eliminate errors where they are very important to catch. The only bad thing which I guess can be reflected in the length of the paper (4 pages) was that several times throughout it the authors said that a larger sample would be needed for better testing. A larger user study should be done to see if the trends in this paper remain true.