Automated detection of Pulmonary nodules in PET/CT images: Improved detection using active contour technique

Abstract

In this study, we proposed a novel pulmonary nodule detection method that introduces active contour technique to detect the nodules deemed “difficult” in previous CAD schemes. The proposed scheme detects pulmonary nodules using both CT and PET images. As for the detection in CT images, the massive region was enhanced using an active contour filter. Subsequently, the PET images were binarized to detect the highly uptake regions. The results were integrated, followed by false positive reduction. In the experiment, we evaluated the proposed method using 104 PET/CT images. As a result, we found that detection performance of proposed method was 6% higher than that of our previous technique.