Artificial Life Models in Lung CTs

A new method for the analysis of 3D medical images is introduced. The algorithm is based on Biological Models of ants known as Artificial Life models. Test images (lung Computed Tomographies) undergo a 3D region growing procedure for the identification of the ribs cage. Active Contour Models are used in order to build a confined area where ants are deployed. The antbased approach, in which steps are allowed in any direction with different probability, allows a kind of tunneling effect for the successful identification of small 3D structures that are not clearly connected to the rest of the tree. The best approach is based on a gradient rule for the release of pheromone. A possible application, as component of a Computer Assisted Detection system for the identification of lung nodules, is the removal of the bronchial and vascular tree from lung CTs.

• Type: Conference
• Year: 2006
• Where: EVOW
• Authors: Sorin Cristian Cheran, Gianfranco Gargano
• Keywords: Active Contour Models, Artificial Intelligence, Artificial Life Models, EVOW 2006, Region Growing Procedure