Automatic Identification and Elastic Properties of Deformed Objects Using their Microscopic Images

In this chapter the state of the art is presented in the domain of automatic identification and classification of bodies on the basis of their deformed images obtained via microscope. The approach is illustrated by means of the case of automatic recognition of third-stage larvae from microscopic images of them in high deformation instances. The introduced methodology incorporates elements of elasticity theory, image processing, curve fitting and clustering methods; a concise presentation of the state of the art in these fields is given. Combining proper elements of these disciplines, we first evaluate the undeformed shape of a parasite given a digital image of a random parasite deformation instance. It is demonstrated that different orientations and deformations of the same parasite give rise to practically the same undeformed shape when the methodology is applied to the corresponding images, thus confirming the consistency of the approach. Next, a pattern recognition method is introduced to classify the unwrapped parasites into four families, with a high success rate. In addition, the methodology presented here is a powerful tool for the exact evaluation of the mechano-elastic properties of bodies from images of their deformation instances. C. Papaodysseus National Technical University of Athens, Greece