作者: Alexander Houston , Angus Craig
DOI: 10.1007/BF00956712
关键词:
摘要: A perceptual experiment was conducted using cardiac phase and amplitude images. Estimates of statistical parameters were derived from the images diagnostic potential human decisions compared. Five methods used to generate 75 gated studies, 39 which classified as pathological. The presented 12 observers experienced in nuclear medicine. rated a five-category scale based on their confidence an abnormality presenting. Circular linear statistics analyse image data, respectively. mean, standard deviation (SD), skewness, kurtosis first term spatial correlation function evaluated region left ventricle. receiver operating characteristic analysis performed both sets data For images, circular SD shown discriminate better between normal abnormal than observers, but no single statistic discriminated well observer for