Extraction automatique d'objets cellulaires en imagerie médicale microscopique: une approche intégrant les contours actifs avec des informations contours et régions

Sophie Schüpp (1) , Abderrahim Elmoataz (1) , Régis Clouard (1) , Daniel Bloyet (2) , Paulette Herlin (3)

1 Equipe Image - Laboratoire GREYC - UMR6072
2 GREYC - Groupe de Recherche en Informatique, Image et Instrumentation de Caen
3 GRECAN - Groupe Régional d'Etudes sur le CANcer

We present a general method of segmentation integrating active contours with information of contours and regions. An automatic and robust initialisation is proposed in order to detect a set of seeds inside of all objects of image. These seeds are classified according to their intensity or colour in several groups. Then, the set of boundaries of these seeds evolves simultaneously under constraints associated to contours and regions in order to localise the final contours of all objects in image. The implementation of multiple evolution of all contours is made by level set method with fast marching method. This method is well suited with a large class of images in medical microscopic imaging. We illustrate the application with quantification of immunostaining. This example of image analysis is representative of problems in quantitative segmentation.

Mots clés: Segmentation; contours actifs; microscopie médicale; histologie; images couleur

Domaines: Traitement des images [eess.IV]

Sophie Schüpp, Abderrahim Elmoataz, Régis Clouard, Daniel Bloyet, Paulette Herlin. Extraction automatique d'objets cellulaires en imagerie médicale microscopique: une approche intégrant les contours actifs avec des informations contours et régions. Vision Interface 2000, 2000, Montreal, Canada. pp.69-75. ⟨hal-00869829⟩