Click here to


Are you sure ?

Yes, do it No, cancel

Integration of High-Velocity Task Object Motion Into Channelized Hotelling Observer Assessment of X-Ray Angiography Systems

A Tao*, K Fetterly , Mayo Clinic, Rochester, MN


(Thursday, 8/2/2018) 10:00 AM - 12:00 PM

Room: Room 207

Purpose: The detrimental effect of motion blur on coronary angiograms is not readily quantifiable with standard image quality metrics. The purpose of this work was to develop theory and methods to incorporate motion blur in the assessment of X-ray angiography system performance using a channelized Hotelling observer model (CHO).

Methods: A rotating platter was used to achieve test object linear velocities ranging from 0.5-30 cm/s (2-120 rpm). Lucite was equally padded above and below the rotating platter to a thickness of 25 cm. Test objects included 160 mgI/mL disks, spheres, and pill-shaped test objects with diameters ranging 0.5-4 mm. The test object phantoms were positioned at isocenter and imaged with cine X-ray technique 96 kV, 0.2 mm copper, 15 fps, and 10 ms pulse width. The detectability index (d’) for each object was estimated using a channelized Hotelling observer which accommodates the circular motion of the test objects. Theory and methods to correct d’ for the detrimental effects of x-ray system non-stationarity were developed.

Results: System non-stationarity reduced d’ of slowly moving objects by ~20%. Following correction for non-stationarity and considering test conditions resulting in d’>1.0, motion was found to affect d’ of all test object shapes similarly. d’ was unaffected by motion up to 2 cm/s. For linear velocity >2 cm/s, d’ decreased with increasing velocity, with d’ of small objects affected to a greater extent than large objects. At linear velocity 30 cm/s d’ was degraded by 50-75%.

Conclusion: A customized CHO model was implemented to estimate the detrimental effect of motion blur on X-ray imaging system performance. d’ degradation due to motion increased with increasing velocity, decreased with decreasing object size, but was largely unaffected by test object shape.


Not Applicable / None Entered.


IM- X-ray: Fluoroscopy, digital angiography, and DSA

Contact Email