Fuzzy Sets in Medical Image Processing
Georg Berks, D. Graf Keyserlingk · Studies in fuzziness and soft computing · 2000
This article describes some aspects of fuzzy methodology and their possible application in the medical community. Fuzzy methods can be arranged according to the hierarchy of medical image processing procedures to allow a convenient comparison with competitive crisp methods. Several examples are explained derived from all the three levels of image processing. Histogram evaluation using fuzzy expective value is a possible application of low level image processing. Structure and function of linguistic variables as aspects of an intermediate level are explained in some detail because of their peculiar ability to cope with uncertainties. A major problem in high level medical image processing is dealt with, when anatomic atlases are superimposed on the highly variable human brain. Finally, a method of comparing distinctive features of anatomical structure is presented.