Transactions on Machine Learning and Data Mining (ISSN: 1865-6781)
Volume 3 - Number 1 - July 2010
Mining 3D Biological Images in order to Construct a Landscape Model of Biological Objects
S. Schmidt, T. Yuan, C. Schatte and P. Perner
Institute of Computer Vision and Applied Computer Sciences, Leipzig, Germany
Abstract
New microscopic image acquistion methods such as the Selective Plane Illumination Microscope (SPIM) allow to produce 3 D images of living organism and to observe processes in deeper tissues of a living organism. With SPIM it is possible to optically cut a probe into several slices. This stack of images must be reconstructed into a 3D object by a visualization technique and automatically analyzed into a 3D Molecular Landscape that allows to describe symbolically the object and its parts and to study the changes in the organism over time. We present the developed reconstruction method based on a cyst imaged by SPIM. The representation of such a 3D molecular landscape is provided and the advantage of the symbolic representation compared to the image data is demonstrated.