Model Observer in Digital Angiography

An important aspect related to angiographic equipment is represented by the high frame-rate of the modality and the performed protocols considered.

Usually employed metric (as Model Observer approaches) are designed to consider only two dimensional images and score is evaluated using only the information taken from single images, without considering any mutual information from contiguous frames. 

It can lead to an underestimation of Performances making these metrics not suitable to be applied with spatio-temporal images at least in condition of high frame rate.

An important temporal characteristic of the human visual system is represented by the noise integration time of the human eye-brain system.

Information derived from cine images are integrated and merged over a fixed period of time; in this way, images noise are globally reduced (increasing detectability) at the cost of a degradation in motion information.

It was estimated as 200 ms.

The noise integration time is than introduced in the statistical method applying throughout a preliminary step in the process.

The Statistical Method was applied on an averaging of frames depending from the visualised fps.

In particular, the averaging process was performed on groups of frames that correspond to the noise integration time of 200 ms, e.g. 3 frames with 15 fps.