A robust lung segmentation algorithm using fuzzy C-means method from HRCT scans

Introduction: Image segmentation has importance role for medical image analysis. Because increasing image-processing methods can provide more reliable information for the best interpretation of diseases. Therefore correct segmentation of the pathological and anatomical structures has critical importance in medical images. Additionally, CT scan interpretation requires lots of time and energy because the large number scans are generated for each patient. Method: For this reasons, in this study a robust and reliable fast algorithm was proposed for the lung segmentation from low-dose HRCT images. The images have been recruited from asthma subjects before and after two months treatment with low-dose limited HRCT scan. HRCT scans were collected in full inspiration for all patients, who were performed with stable and mild moderate asthma. Fuzzy Cmeans method, which is based on image histogram, was used in this study. Algorithm was evaluated on 20 asthma patients. Conclusion: The accuracy is approximately 93.7% and it takes almost a second for each slice. In future studies we are planning to study airway segmentation and measure airway values before and after treatment.