The practical application of signal detection theory to image quality assessment in x-ray image intensifier-TV fluoroscopy

This paper applies a published version of signal detection theory to x-ray image intensifier fluoroscopy data and compares the results with more conve ntional subjective image quality measures. An eight-bit digital framestore was used to acquire temporally contiguous frames of fluoroscopy data from w hich the modulation transfer function (MTF(u)) and noise power spectrum wer e established. These parameters were then combined to give detective quantu m efficiency (DQE(u)) and used in conjunction with signal detection theory to calculate contrast-detail performance. DQE(u) was found to lie between 0 .1 and 0.5 for a range of fluoroscopy systems. Two separate image quality experiments were then performed in order to asse ss the correspondence between the objective acid subjective methods. First, image quality for a given fluoroscopy system was studied as a function of doserate using objective parameters and a standard subjective contrast-deta il method. Following this, the two approaches were used to assess three dif ferent fluoroscopy units. Agreement between objective and subjective method s was good: doserate changes were modelled correctly while both methods ran ked the three systems consistently.