| |
      |
|
Schematic of a general framework for a model observer that includes a variety of components reflecting the optics, channels in the primary visual cortex, non-linear decision rules including channel inhibition (from Zhang et al., IEEE TMI 2006)
There is a strong rationale to develop computer models that allow for fast and reliable evaluation and optimization of medical image quality. Human psychophysical studies are time consuming and often times simply not viable for optimization when parameters of image acquisition/display suffer combinatorial explosions.Our work in this area has concentrated on both: 1) advancing model observers by testing the proposed models with new components and extending the models to more realistic scenarios;
2) illustrating how model observers can be used to evaluate and optimize medical image quality.
Our work has concentrated on x-ray coronary angiography and mammography and more recently breast ultrasound and breast computer tomography (CT).Our theoretical work on model observers include:
|
|
