ABSTRACT
Image processing provides tools for image enhancement and extraction of information from images. Modification of the colour table allows for amplification of the contrast in certain pixel-value intervals while the contrast is suppressed in other intervals. This can in turn be used to selectively enhance the displayed image in intervals of special importance to the current problem. One of the most important conflicts in imaging is the trade-off between a high signal-to-noise-ratio (SNR) and a good spatial resolution. Low-pass filtering can be used to improve the SNR at the expense of spatial resolution. Likewise, high-pass filtering can be used to enhance edges but tends to be noise sensitive. Another important kind of operations is spatial transformations, for example, scaling, translation, and rotation. The discrete nature of digital images makes the application of such operations dependent on the use of interpolation. There are different interpolation methods that vary in accuracy and computational burden. Lastly, many applications are dependent on image segmentation. The most common segmentation method is manual delineation, but there are a number of automatic and semi-automatic methods that may be used to save time and improve reproducibility. While the presentation of the subject is by no means complete, the hope is that it will serve as an introduction to the most important image-processing techniques used in nuclear medicine.
