Fully automated liver segmentation from SPIR image series Evgin Göçeri a,n , Metin N. Gürcan b ,Oğuz Dicle c a Department of Computer Engineering, Pamukkale University, Denizli, Turkey b Department of Biomedical Informatics, The Ohio State University, Columbus, OH, USA c Department of Radiology, Faculty of Medicine, Dokuz Eylul University, Narlıdere, Izmir, Turkey article info Article history: Received 11 June 2014 Accepted 10 August 2014 Keywords: Variational level set SPIR Liver segmentation Signed pressure force function Active contour abstract Accurate liver segmentation is an important component of surgery planning for liver transplantation, which enables patients with liver disease a chance to survive. Spectral pre-saturation inversion recovery (SPIR) image sequences are useful for liver vessel segmentation because vascular structures in the liver are clearly visible in these sequences. Although level-set based segmentation techniques are frequently used in liver segmentation due to their flexibility to adapt to different problems by incorporating prior knowledge, the need to initialize the contours on each slice is a common drawback of such techniques. In this paper, we present a fully automated variational level set approach for liver segmentation from SPIR image sequences. Our approach is designed to be efficient while achieving high accuracy. The efficiency is achieved by (1) automatically defining an initial contour for each slice, and (2) automatically computing weight values of each term in the applied energy functional at each iteration during evolution. Automated detection and exclusion of spurious structures (e.g. cysts and other bright white regions on the skin) in the pre-processing stage increases the accuracy and robustness. We also present a novel approach to reduce computational cost by employing binary regularization of level set function. A signed pressure force function controls the evolution of the active contour. The method was applied to ten data sets. In each image, the performance of the algorithm was measured using the receiver operating characteristics method in terms of accuracy, sensitivity and specificity. The accuracy of the proposed method was 96%. Quantitative analyses of results indicate that the proposed method can accurately, efficiently and consistently segment liver images. & 2014 Elsevier Ltd. All rights reserved. 1. Introduction Liver is an essential organ with several vital functions such as protein synthesis and detoxification. Additionally, it regulates biochemical reactions that include the synthesis or breakdown of complex and small molecules and produces bile, which is an alkaline compound aids in digestion. Yet, no technique or device can compensate for the absence of the liver. The only available option is liver transplantation, which is a major and risky surgery. Although transplantation from cadavers utilized to be the first choice, transplantation from living donors has become a choice of treatment due to the shortage of cadaver donation in recent years [1]. Before the surgical procedure, the livers that belong to the living donor and the recipient are evaluated: to identify the liver region, to determine the size mismatch, to measure liver volume, and to analyze the vascular structure. Knowledge obtained by the evaluation is needed to decide whether the donor and recipient is a good match and when the transplantation should be performed. The success of the surgical operation and reduction of complica- tions (which may occur during or after the operation) to minimum level depends on accuracy of anatomic information of the portal and hepatic veins, compatibility of these vessels and liver volume. Therefore, precise measurements and analysis of liver and vessels that requires accurate liver segmentation from all image slices have vital importance for liver transplantation at pre-evaluation stage [2]. Several automatic and semi-automatic liver segmentation methods from Computed Tomography (CT) [3] and from Magnetic Resonance (MR) images [4–11] have been proposed to overcome problems of manual liver segmentation. An important problem of manual segmentation is that liver boundaries can be identified differently by different radiologists and even by the same radi- ologist at a different time (i.e. inter and intra-reader variability). Thus, segmentation results depend on experience and skills of radiologists. Also, it is very time consuming and tedious task because of the high number of image slices and data sets. Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/cbm Computers in Biology and Medicine http://dx.doi.org/10.1016/j.compbiomed.2014.08.009 0010-4825/& 2014 Elsevier Ltd. All rights reserved. n Corresponding author. E-mail address: egoceri@pau.edu.tr (E. Göçeri). Computers in Biology and Medicine 53 (2014) 265–278