This paper describes the sequence of construction of a cell nuclei classification model by the analysis, the characterization and the classification of shape and texture. We describe first the elaboration of dedicated shape indexes and second the construction of the associated classification submodel. Then we present a new method of texture characterization, based on the construction and the analysis of statistical matrices encoding the texture. The various characterization techniques developed in this paper are systematically compared to previous approaches. In particular, we paid special attention to the results obtained by a versatile classification method using a large range of descriptors dedicated to the characterization of shapes and textures. Finally, the last classifier built with our methods achieved 88% of classification out of the 94% possible.

In this paper, we present an approach to extract feature lines on a 3D mesh. The main problem of classical methods is the difficulty to obtain connectivity of resulting lines. In order to sort out this recurrent problem, we propose an algorithm based on the following ideas. The first one is to label candidate points: they characterize the relevant areas corresponding to the expected features. Our second idea is to unfold this binary flagged mesh into a 2D space in order to use well-known morphological operators. By mapping the resulting output back on the mesh, we get relevant and connected feature lines. We then apply this technique to geological data scans to extract fractures and bench limits.