Automatic 3D seed location and orientation detection in CT image for prostate brachytherapy

In prostate brachytherapy, the analysis of the 3D pose information of each individual implanted seed is one of the critical issues for dose calculation and procedure quality assessment. This paper addresses the development of an automatic image processing solution for the separation, localization and 3D orientation estimation of prostate seeds. This solution combines an initial detection of a set of seed candidates in CT images (using a thresholding and connected component method) with an orientation estimation using principal components analysis (PCA). The main originality of the work is the ability to classify the detected objects based on a priori intensity and volume information and to separate groups of seeds using a modified k-means method. Experiments were carried out on CT images of a phantom and a patient aiming to compare the proposed solution with manual segmentation or other previous work in terms of detection performance and calculation time.