Abstract
Accurate localization and analyses of functional liver segments are crucial in devising various surgical procedures, including hepatectomy. To this end, they require the extraction of a liver from computed tomography, and then the identification of resection correspondence between individuals. The first part is usually impeded by inherent deficiencies, as present in medical images, and vast anatomical variations across subjects. While the model-based approach is found viable to tackle both issues, it is often undermined by an inadequate number of labeled samples, to capture all plausible variations. To address segmentation problems by balancing between accuracy, resource consumption, and data availability, this paper presents an efficient method for liver segmentation based on a graph-cut algorithm. One of its main novelties is the incorporation of a feature preserving a metric for boundary separation. Intuitive anatomical constraints are imposed to ensure valid extraction. The second part involves the symmetric conformal parameterization of the extracted liver surface onto a genus-0 domain. Provided with a few landmarks specified on two livers, we demonstrated that, by using a modified Beltrami differential, not only could they be non-rigidly registered, but also the hepatectomy on one liver could be envisioned on another. The merits of the proposed scheme were elucidated by both visual and numerical assessments on a standard MICCAI SLIVER07 dataset.
Highlights
Subject-specific modeling of 3D organs plays a crucial part in computerized diagnosis and therapeutic intervention
Numerous techniques for segmenting a 3D liver from computed tomography (CT) have been proposed. These techniques relied on different features being extracted from imaging data, and can be generally classified into those based on thresholding [11,15], region growing (RG) [13,16,17], graph processing [5,18,19,20,21,22], machine learning [23,24], level-set [25], and deformable model [26]
The shape (kd1eerf−ninegθel((dippsa))ds)=|ei∇nfif12nE(aepqtdu)a|nab2t yioannN(5epN)llaipfpnx22sdf−ex,∇fwfyyf2dh(xpodds)xyedisyma anjo+irmπa2angde axes are denoted by s1 and s2, respectively, as expressed in Equations (8a) and (8b): wcohoerrdeinfxataensd, rfeyssa1p(reepc)ttih=vee11lys+.tTochrr(depe)lro⁄cλpaalrntieailgdhebroivrhaotiovdesoof fptihsediemnaogteed, f,aws Nit(hp8.ar)espect to x and y thse2(apn)is=ot(r1op−icgm(pe)a)ssu1(rpe)and corner strength, used to estim(a8teb)the kernel shape where r is the issugpipvoernt froardeiaucshopf iaxekl,eprn, erel,spwehcitcivheilsy,saest ftoolltohwrese: in this paper, and λ is a factor controlling the importance of the imaging feature
Summary
Subject-specific modeling of 3D organs plays a crucial part in computerized diagnosis and therapeutic intervention. Numerous techniques for segmenting a 3D liver from CT have been proposed These techniques relied on different features being extracted from imaging data, and can be generally classified into those based on thresholding [11,15], region growing (RG) [13,16,17], graph processing [5,18,19,20,21,22], machine learning [23,24], level-set [25], and deformable model [26]. The method was further enhanced by employing max-flow/min-cuts algorithms in minimization [30], and has since been widely received by many studies Taking advantage of these state-ofthe-art algorithms, Chen et al [21] proposed combining the active appearance model, live wire, and GC for segmentation of the liver, kidneys and spleen. Simple linear iterative clustering (SLIC) and super voxel-based GC Largest connected component (LCC) and Bayesian model
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
