Journal of Computational Science 25 (2018) 376–387 Contents lists available at ScienceDirect Journal of Computational Science j ourna l h om epage: www.elsevier.com/locate/jocs Computational intelligence optimization approach based on particle swarm optimizer and neutrosophic set for abdominal CT liver tumor segmentation Ahmed M. Anter a,c,∗ , Aboul Ella Hassenian b,c a Faculty of Computers and Information, Beni-Suef University, Benisuef, Egypt b Faculty of Computers and Information, Cairo University, Cairo, Egypt c Scientific Research Group in Egypt, (SRGE), Egypt 1 a r t i c l e i n f o Article history: Received 31 December 2016 Received in revised form 28 November 2017 Accepted 9 January 2018 Available online 31 January 2018 Keywords: Meta-heuristic Particle swarm optimization Segmentation Neutrosophic set a b s t r a c t In this paper, an improved segmentation approach for abdominal CT liver tumor based on neutrosophic sets (NS), particle swarm optimization (PSO), and fast fuzzy C-mean algorithm (FFCM) is proposed. To increase the contrast of the CT liver image, the intensity values and high frequencies of the original images were removed and adjusted firstly using median filter approach. It is followed by transforming the abdominal CT image to NS domain, which is described using three subsets namely; percentage of truth T, percentage of falsity F, and percentage of indeterminacy I. The entropy is used to evaluate indeterminacy in NS domain. Then, the NS image is passed to optimized FFCM using PSO to enhance, optimize clusters results and segment liver from abdominal CT. Then, these segmented livers passed to PSOFCM technique to cluster and segment tumors. The experimental results obtained based on the analysis of variance (ANOVA) technique, Jaccard Index and Dice Coefficient measures show that, the overall accuracy offered by neutrosophic sets is accurate, less time consuming and less sensitive to noise and performs well on non-uniform CT images. © 2018 Elsevier B.V. All rights reserved. 1. Introduction Segmentation is a critical and essential process and is one of the most difficult tasks in image processing. Automatic segmen- tation of CT liver tumor is a very challenging task, due to various factors, such as the low-level contrast and blurry edged images, irregularity in the liver shape and size between the patients and the similarity with other organs of almost same intensity like spleen and stomach. Also, liver parenchyma is stretched over 150 slices in a CT image and different from patients, indefinite shape of the lesions and low intensity contrast between lesions and similar to those of nearby tissues make automatic liver and lesions segmenta- tion difficult [1,2]. Among various image segmentation techniques, traditional segmentation methods have certain drawbacks, which cannot be used for accurate result and time computation. ∗ Corresponding author at: Faculty of Computers and Information, Beni-Suef Uni- versity, Benisuef, Egypt. E-mail address: sw anter@yahoo.com (A.M. Anter). 1 http://www.egyptscience.net. Fuzzy theory has been applied to image segmentation, which retains more information than that of the hard segmentation meth- ods. Fuzzy C-means (FCM) is a fuzzy clustering method allowing a piece of data to belong to two or more clusters. The FCM algo- rithm obtains segmentation results by using fuzzy classification [3]. In some applications such as CAD systems, we should con- sider not only the truth and falsity membership, but also we want the indeterminacy membership. It is hard for classical fuzzy set to solve such problems [3,4]. As a generalization of fuzzy logic, neu- trosophic logic introduces a percentage of ‘indeterminacy’ due to unexpected parameters hidden in some propositions and carries more information than fuzzy logic [5]. Many problems in medical images have been solved by consid- ering bio-inspired meta-heuristic optimization algorithms such as Social Spider Optimization (SSO), Ant Colony Optimization (ACO), Crow Search Optimization (CSO), and particle swarm optimization (PSO). Computational bio-inspired algorithms have been used in situations where conventional techniques cannot find a satisfactory solution or they take too much time to find the solution [42]. There- fore, this paper introduces a very powerful optimization method, both in terms of speed and optimal convergence, which can be con- https://doi.org/10.1016/j.jocs.2018.01.003 1877-7503/© 2018 Elsevier B.V. All rights reserved.