A new processing sequence to assess airways using 3D CT-scan

Accurate measurement of physiological parameters such as the bronchial lumen and bronchial wall areas helps in the diagnosis of lung diseases. Algorithms allowing such measures are the result of complex processing sequences composed of (i) bronchial volume extraction (ii) bronchial volume skeletonization (iii) wall extraction. The proposed sequences however still show weaknesses in terms of precision, extent and automation. In this paper we address various problems encountered at each step of the processing sequence and propose a new sequence for measuring lumen on bronchial section. Finally, this sequence is tested on two datasets of 10 anonymous exams¹ collected from 10 different healthy volunteers and 10 different patients suffering from asthma that were attended by different practitioners.