Intrathoracic Airway Tree Segmentation from CT Images Using a Fuzzy Connectivity Method

Document Type : Original Paper


1 Master of Science in Biomedical Engineering, Department of Biomedical Systems & Medical Physics, Tehran University of Medical Sciences

2 Associate Professor in Biomedical Engineering, Biomedical Systems & Medical Physics Dept., Tehran University of Medical Sciences & Research Center for Science and Technology in Medicine, Tehran, Iran.

3 Assistant Professor in Biomedical Engineering, Electrical Engineering Dept., Sharif University of Technology, Tehran, Iran.

4 Associate Professor in Biomedical Engineering, Electrical Engineering Dept., Ryerson University, Toronto, Canada.

5 Assistant Professor, Pneumologist Consultant, Iran University of Medical Sciences, Internal Medicine Group, Tehran, Iran.


Introduction: Virtual bronchoscopy is a reliable and efficient diagnostic method for primary symptoms of lung cancer. The segmentation of airways from CT images is a critical step for numerous virtual bronchoscopy applications.
Materials and Methods: To overcome the limitations of the fuzzy connectedness method, the proposed technique, called fuzzy connectivity - fuzzy C-mean (FC-FCM), utilized the FCM algorithm. Then, hanging-togetherness of pixels was handled by employing a spatial membership function. Another problem in airway segmentation that had to be overcome was the leakage into the extra-luminal regions due to the thinness of the airway walls during the process of segmentation.
Results:   The result shows an accuracy of 92.92% obtained for segmentation of the airway tree up to the fourth generation.
Conclusion:  We have presented a new segmentation method that is not only robust regarding the leakage problem but also functions more efficiently than the traditional FC method. 


Main Subjects