Airway tree segmentation based on CT images of patients with acute respiratory distress syndrome (ARDS)
Citas bibliográficas
Enlace de Referencia
Autores
Director
Autor corporativo
Recolector de datos
Otros/Desconocido
Director audiovisual
Editor/Compilador
Fecha
Resumen
The purpose of this project was to study the influence of the airway volume over the lung volume for acute respiratory distress syndrome (ARDS). For achieving this, a method of region growing and Variational Region Growing (VRG) adapted for airway segmentation were tested and voxels of segmentation results were classified into regions depending on the aeration level as over aerated, normally aerated, poorly aerated or non-aerated. Finally, results of each region for region growing method were compared against results for VRG method using Pressure-volume curves (PV-curves). Tests were carried out over Computer Tomography (CT) volumes for 8 pigs with induced ARDS at various aeration conditions. A total of 37 CT images were analyzed for pig. The results indicated that VRG method goes to the 4 or 5 generation of airway and segments 14% more of bronchi in pathologic lungs than a region growing algorithm. It was found that most of the bronchi reside in the over aerated class and they represent 2% of the lung area in the pig. Despite those results, it is recommended further research to improve accuracy of airway detection and a technique for results validation and branch counting.