Corrections of errors in densitometry under simulated in vivo conditions

Errors in densitometry due to distortion, scattering, and beam hardening have been eliminated under simulated in vivo conditions. The method was tested by means of an anthropomorphic phantom. The results reveal good agreement between the evaluated relative lumina and the real vessel relations. There is a linear relationship between the integrated profile of gray values and the real cross section. With this procedure the measured cross section contains a mean error of about 5%-7%.<
>