POSTER: Shape Indexes: from 2D to 3D. Application to Cells' Nuclei Classification

Thibault,G., Devic,C., Fertil,B., Mari,J.L., Sequeira,J.

In this paper, we present a study on the characterization and the classification of binary digital shapes. This study is performed using a set of values obtained by the computation of "shape indexes". To get these indexes, we extract a set of data called "measures" from the 2D shape, like for example a shape's surface and perimeter. Then these measures are used as parameters of a function bringing a real value which gives information about geometrical and morphological features on the shape to analyze. These 2D shape indexes are then extended to 3D to provide an original method of texture characterization. A model of shape and texture characterizations is built based on this numerical information. It is used to classify cells nuclei in order to diagnose patients affected by the Progeria disease.