Feature Extraction for Cortical Sulci Identification

Nicolas Royackkers (1) , Houssam Fawal (1) , Michel Desvignes (1) , Marinette Revenu (1) , J.-M. Travère (2)

1 Equipe Image - Laboratoire GREYC - UMR6072
2 Cyceron - GIP Cyceron

The use of PET in quantitative measurement of brain activity requires the superimposition of some anatomical data coming from other sources, like MRI. In the frame of structural and anatomical data matching for a great number of patients, we developed a method for the automatic identification of cortical sulci on 3D MR images. The knowledge used is automatically extracted from a database containing a few pictures where sulci have been previously recognized. Proportional correction mechanisms, based on Talairach's grid, are proposed. They intend to adapt sulci statistical models to the particular features of any brain, in order to make the recognition easier. Our identification method is efficient and robust for the superficial part of six major sulci.

Domaines: Traitement des images [eess.IV]

Nicolas Royackkers, Houssam Fawal, Michel Desvignes, Marinette Revenu, J.-M. Travère. Feature Extraction for Cortical Sulci Identification. 9th SCIA, 1995, Uppsala, Sweden. pp.1147-1154. ⟨hal-00963711⟩