Autors: KalchevaYovkova, E. P., Goranova, M. E.
Title: An Approach for Automatic Rib Cage Segmentation and Evaluation from Axial CT Images
Keywords: Computed tomographic (CT) images, abdominal bone structures, morphological image processing, image segmentation, evaluation

Abstract: Preliminary results of an automatic approach for bone structures segmentation from axial computed tomographic (CT) images are presented. The proposed method combines the morphological image processing techniques with prior anatomical knowledge of human abdomen. The approach is tested on CT data sets from 15 patients; each data set consists between 95 and 150 images. The results of the segmentation of the spine and the ribs are assessed quantitatively by comparing with the "ground truth" - a manual segmentation performed by a radiologist. The average values of specificity, sensitivity and accuracy are relatively high (above 0,96) and mean value of Dice coefficient is about 0,6.



    International Scientific Conference Computer Science’2015, pp. 262, 2015, Albania, ISBN 978-619-167-177-9

    Вид: публикация в международен форум