Comparison of two methods for evaluating image quality of chest radiographs

摘要: A set of 15 analog chest images was digitized with a high performance scanner and manipulated in terms noise resolution to yield three sets images; the original plus two different properties. These were evaluated Visual Grading Analysis (VGA) where observer rates visibility certain normal anatomical structures (as described by European Quality Criteria) compared reference image. One non-manipulated used an ROC-related method – Free response Forced Error experiment. Simulated lesions contrast size randomly superimposed on 50 copies this The then same way as for VGA. All observations done group seven expert radiologists from six countries all printed back film before evaluation. ranking image manipulations methods. Although strong correlation cannot be predicted only images, it is encouraging that simpler VGA study some cases might replacement more time- effort-consuming ROC analysis. (Less)