Snakes based tracking and texture analysis of microscopic images

Abstract

Snake based tracking over a textured potential is applied to recover the 3D structure of a complex organ, the eye, from a sequence of slices. Grey level images are first analysed by a probabilistic method to derive a suitable potential. Then, the snake is initialised by hand and, finally, its energy function is minimised. Temporal discontinuity problems due to the nature of the images (microscopic slices tinted and taken by hand) are solved.