Robust and fully automated segmentation of mandible from CT scans

Abstract

Mandible bone segmentation from computed tomography (CT) scans is challenging due to mandible's structural irregularities, complex shape patterns, and lack of contrast in joints. Furthermore, connections of teeth to mandible and mandible to remaining parts of the skull make it extremely difficult to identify mandible boundary automatically. This study addresses these challenges by proposing a novel framework where we define the segmentation as two complementary tasks: recognition and delineation. For recognition, we use random forest regression to localize mandible in 3D. For delineation, we propose to use 3D gradient-based fuzzy connectedness (FC) image segmentation algorithm, operating on the recognized mandible sub-volume. Despite heavy CT artifacts and dental fillings, consisting half of the CT image data in our experiments, we have achieved highly accurate detection and delineation results. Specifically, detection accuracy more than 96% (measured by union of intersection (UoI)), the delineation accuracy of 91% (measured by dice similarity coefficient), and less than 1 mm in shape mismatch (Hausdorff Distance) were found.

Automatic segmentation of the mandible from computed tomography scans for 3D virtual surgical planning using the convolutional neural network

“…In particular, the testing time costs increase to 10 minutes in the 109 CT scans. We obtain similar accuracies compared to the work of the traditional methods (Mannion-Haworth et al 2015) (Torosdagli et al 2017) but with superiority in efficiency and fully-automatic nature. Furthermore, such a design of the system enables the consideration of similarity and structural continuity of the mandible from different planes.…”

“…Conventional manual segmentation of the mandible in CT scans leads to a tedious procedure in the clinical practice (Huff, Ludwig & Zuniga 2018). Moreover, the structural complexity of mandibles and the considerable human rater variability make the segmentation of mandibles in CT scans challenging (Torosdagli, Liberton, Verma, Sincan, Lee, Pattanaik & Bagci 2017). Manual segmentation also has limited reproducibility and is very time-consuming.…”

“…Conventional approaches for mandible segmentation typically use pixel-based or model-based methods. In the past decade, some traditional semi-automatic and automatic methods have been developed to segment the mandible in CT scans (Gollmer & Buzug 2012) (Torosdagli et al 2017) (Chuang, Doherty, Adluru, Chung & Vorperian 2017) (Abdi, Kasaei & Mehdizadeh 2015) )(Mannion-Haworth, Bowes, Ashman, Guillard, Brett & Vincent 2015 . A statistical shape model for mandible segmentation was presented by Gollmer and Buzug in 2012 (Gollmer & Buzug 2012).…”

“…A statistical shape model for mandible segmentation was presented by Gollmer and Buzug in 2012 (Gollmer & Buzug 2012). Torosdagli et al proposed a 3D gradient-based fuzzy connectedness algorithm to segment mandibles (Torosdagli et al 2017). A registration-based semiautomatic mandible segmentation technique was proposed by Chuang et al in 2017(Chuang et al 2017 with the atlases at the global level, which allowed multi-atlas-based segmentations and correlation-based label fusion to be performed at the local level.…”

CT image segmentation of bone for medical additive manufacturing using a convolutional neural network

“…Moreover, the mean DSC found in the present study is comparable to the results reported by Fu et al [50] (2017) who proposed an atlas-based method and achieved a mean DSC of 0.94 ± 0.01 when segmenting the mandible. Torosdagli et al [51] (2017) developed a 3D gradient-based fuzzy connectedness method for the segmentation of the mandible and reported a DSC of 0.91. Furthermore, the DSCs found in the present study are higher than those reported by Jafarian et al [14] (2014) and Ghadimi et al [52] (2016), who segmented cranial bones of neonates using a level-set method and achieved mean DSCs of 0.87 and 0.81, respectively.…”