Evaluation of a psychophysical model of observer performance in the detection of low contrast radiographic patterns

G. Pedroli, A. Crespi, C. Ghezzi, G. Montanari, N. Paruccini, E. Salvini, M. Tagliabue

Research output: Contribution to journalArticle

Abstract

The purpose of this study was to evaluate the human observer detection performance in normal viewing conditions of radiographic images. The visual detectability of circular objects in radiographic images obtained with a digital radiography system was measured by means of a sixteen-alternative forced-choice (16AFC) method. Physical image quality indices were calculated according to three different models of the detection process (amplitude model, quadratic model and statistical decision theory model) by taking into account system MTF, noise Wiener spectrum, visual response function and the noise component internal to the observer. The results of this work indicate that the statistical decision theory model including the noise component internal to the observer but not the visual response function most closely predicts the observer performance in the detection of low contrast radiographic patterns.

Original languageEnglish
Pages (from-to)339-343
Number of pages5
JournalPhysica Medica
Volume13
Issue numberSUPPL. 1
Publication statusPublished - 1997

Keywords

  • Image quality indices
  • Psychophysical models
  • Signal- to- noise ratio

ASJC Scopus subject areas

  • Biophysics
  • Physics and Astronomy(all)

Fingerprint Dive into the research topics of 'Evaluation of a psychophysical model of observer performance in the detection of low contrast radiographic patterns'. Together they form a unique fingerprint.

Cite this