(Monday, 7/30/2018) 9:30 AM - 10:00 AM
Room: Exhibit Hall | Forum 9
Purpose: In this study, we propose a unique fully automated method for lumen segmentation and normal/abnormal lumen classification.
Methods: To detect the lumen morphology, the proposed algorithm transforms the polar into Cartesian coordinate of IVOCT images. The guide wire shadow region and the stent struts can be detected through the pattern analysis. The guide wire shadow region is interpolated, and the stent struts shadow region is filled with high pixel values. Active contour model is used to detect irregular lumen morphology. In total, 92 features were extracted to classify normal/abnormal lumen. The lumen classification method is a combination of supervised machine learning algorithms and feature filtering method.
Results: Abnormal IVOCT images including irregular lumen can be detected and classified in a few seconds. Compared with the manual segmentation result, the dice similarity is 97.2%. The accuracy of normal/abnormal lumen classification is 98.2%. The results can lead to understanding of overall vascular status and help to determine cardiovascular diagnosis.
Conclusion: The proposed algorithm can help clinicians to get more intuitive understand the condition of the stented segment. In addition, the protrusion and thrombus can be easily detected through the proposed algorithm.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
