Preliminary 3D performance evaluation on automatic lung segmentation for interstitial lung disease using high resolution
Computed tomography, Chuen Rue Ng, Joel Chia Ming Than, Norliza Mohd Noor, Omar Mohd Rijal, Rosminah M. Kassim, Ashari Yunus · 2017
Lung disease has been reported by World Health Organization (WHO) to have caused 2.77% of total death in Malaysia and is the 6thmain cause of death which includes interstitial lung disease (ILD). ILD causes alteration and fibrosis in the lungs leading to breathing complications and can be diagnosed and analyzed using high resolution computed tomography (HRCT). Previous study has developed a robust CAD-based system that segments lungs on five levels determined by radiologists. This study is an extension of previous study and focus on the performance evaluation in three dimension (3D) with the segmentation done on all slices of the HRCT lung data. The automated segmentation system shows better results when performance evaluation is done in 3D with Dice's similarity coefficient at 98.32%. Moreover, 3D rendering of lungs and its volume can aid radiologist and medical personnel in diagnosis and analysis of lung diseases.