On fuzzy enhancement of megavoltage images in radiation therapy

The megavoltage electronic portal imaging is a very important task for cancer treatment in radiation therapy. The megavoltage images (MVIs), however, are very poor in quality because of imaging physics. For a reliable patient set-up verification we need better in-treatment images to track relevant features. In recent years, some researchers have applied the concept of fuzzy set theory to develop new image enhancement algorithms. In this work we investigate three existing fuzzy approaches to enhancement of MVIs. We develop also a simple fuzzy rule based approach to image enhancement. Further, we demonstrate the results of the locally adaptive implementation of these methods. Finally, a comparison of the performance of the investigated algorithms is given by the use of some measures of fuzziness. Some general notes on these methods is also given