Related Links

Go Back

Assessment of image quality through objective tests

Introduction

The assessment of image quality using objective tests provides a tool not only for evaluating image quality, but for communication with others, including manufacturers, regarding the quality of medical images.  Some tests are modality specific and some results vary significantly from one modality to another.  For example, the resolution for a typical radiographic imaging system is specified in cycles per millimeter (in cycles per centimeter for computed tomography) and through the use of modulation transfer function (MTF).  It is also essential to have an understanding of what the image quality measurement results should be for a specific modality.

Important Principles

The medical physicist must be able to select the appropriate equipment, software, and measurement techniques for evaluating image quality.  Image quality requirements vary depending on modality and, often, on the clinical task.  Consequently, the physicist must have knowledge of what resolution, contrast, noise level, etc. can be expected from this modality and how that applies to clinical tasks.

Image processing plays a significant role in medical imaging today.  Almost every digital image is processed in some way.  It is important to understand how image processing impacts image quality characteristics.  For example, the amount of noise in an image can be reduced by smoothing the image but a smoothing filter reduces image resolution. Likewise, data compression is essential considering the volume of data produced in medical imaging facilities. However, compression can affect the quality of the image, depending on the level of compression selected.

Information is available in the literature regarding what levels of image quality should be anticipated for specific modalities.  This information, along with the results of previous equipment evaluations, can be used as a benchmark for comparison purposes.

Introduction to References

There are several good references on the objective evaluation of image quality. The website by Sprawls is excellent and, of course, free. Bushberg has become the standard text in the North America while Dowsett has become the standard in the U.K. and Europe.