Ranjit Singh Sandhu MD, Presenter: Nothing to Disclose

Brian Jin MD, Abstract Co-Author: Nothing to Disclose

Richard S. Ha MD, Abstract Co-Author: Nothing to Disclose

Ralph Thomas Wynn MD, Abstract Co-Author: Nothing to Disclose

CAD should not be seen as a "black box." Understanding how CAD analyses images and determines relevant outputs for the radiologist is fundamental for optimal diagnostic use as well as improving quality in radiology. Through participation in this exhibit, the audience will learn and demonstrate understanding of the goals, roles, applications and future directions of CAD.

Introduction to CAD and its potential role in solving issues faced by the field of radiology today. CAD goals (i.e. what are the characteristics of an optimal CAD system?). Overview of CAD image analysis. Image preprocessing along with a discussion about image standardization. Image segmentation. Candidate detection. Feature extraction, candidate analysis/classification. Introductions to vector space and machine learning. System output (i.e. what is CAD trying to tell you as the radiologist and why?). Discussion of first reader vs. second reader and the radiologist's relationship with CAD. Current and future applications. Current limitations. Throughout and at the end of the exhibit, interactive quiz questions will be presented to reinforce concepts and emphasize key points.

http://abstract.rsna.org/uploads/2014/14006299/14006299_qfih.pdf

Sandhu, R, Jin, B, Ha, R, Wynn, R, Computer-aided Diagnosis and You (the Radiologist).  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14006299.html