The availability of systems for digitizing biomedical thin sections has led to a large number of techniques for the computer-aided analysis of these images. Automatic analysis of these sections has been hampered by threshold dependent techniques. This is probably due to the complexity of biological images and the difficulty of adequately parameterizing color characteristics of various stains. Repetition of the results of analysis is extremely important in this class of images, and dependence on thresholds will have an adverse effect on this goal. Mathematical morphology is the study of characteristics of shape and interrelationships between shapes. It is based on set theory operations originally examined by Minkowski, Matheron and Serra. Thus far, biomedical analysis has been limited to black and white or gray scale images. Our short term goal is the development of an imaging algebra based on mathematical morphology operations for the analysis of full color biomedical thin sections. The set theory basis is derived from our previously developed computer vision system design. Long term goals include the study of new mathematical morphology operations for biomedical image analysis, and the study of various multiple valued logic systems for suitability as a basis.
