Regular Paper Leukocyte segmentation in tissue images using differential evolution algorithm Mukesh Saraswat n , K.V. Arya, Harish Sharma ABV-Indian Institute of Information Technology and Management, Gwalior, India article info Article history: Received 16 June 2012 Received in revised form 5 December 2012 Accepted 16 February 2013 Available online 28 February 2013 Keywords: Nuclei Image segmentation Multi-level threshold Differential evolutionary algorithms abstract An automatic segmentation of leukocytes can assist pharmaceutical companies to take decisions in the discovery of drugs and encourages for development of automated leukocyte recognition system. Segmentation of leukocytes in tissue images is a complex process due to the presence of various noise effects, large variability in the images and shape of the nuclei. Surprisingly, rare efforts have been made to automate the segmentation of leukocytes in various disease models on hematoxylin and eosin (H&E) stained tissue images. The present work proposes a novel strategy based on differential evolution (DE) algorithm to segment the leukocytes from the images of mice skin sections stained with H&E staining and acquired at 40  magnification. The proposed strategy is a first inline report used in such type of image database. Further, the proposed strategy is compared with well-known segmentation algorithms. The results show that the proposed strategy outperforms the traditional image segmentation techniques. & 2013 Elsevier B.V. All rights reserved. 1. Introduction Inflammation is a complex protective reaction where injurious agents are either destroyed, diluted or walled-off [1]. Inflamma- tion nearly occurs in every known disease at some time in their course. A controlled inflammatory reaction is a part of protective mechanism; however, in many instances identification and removal of inflammation are not achieved. This reaction is very deleterious. A number of anti-inflammatory drugs are available [2] but due to their reported harmful side effects in recent posts, some of them are banned in market for further use (e.g. Rofe- coxib). Thus still there is a challenge for scientists to discover anti-inflammatory drugs with minimal side effects. Inflammation has two phases: exudative (acute inflammation phase) and cellular (sub-acute or chronic phase) [1]. A variety of instruments are available to aid automation of acute inflamma- tory reactions for reducing human subjectivity and workload. However, rare work have been done in the quantification of inflammatory cells in the tissue images due to their wide natural biological variability. Recently, we have noticed articles on chicken skin model [2], mice skin model [3], air pouch model [4], etc. where inflammatory cells (leukocytes) were quantified manually. Manual counting of leukocytes is a time consuming process and requires a trained dedicated pathologist. The number of variables such as chances of biasness and variations in staining characteristics pose challenges for the research scientists in counting leukocytes manually [5,6]. Therefore, to reduce these problems, there has been a growing interest in developing tools for quantification of leukocytes using image processing techni- ques. Image processing techniques are widely used in the field of medical imaging, computer vision, security, etc. [7–9]. Therefore, it is interesting to use the image processing techniques for quantifying the leukocytes. Some useful algorithms have already been developed for separation of leukocytes in blood smears [10,11] but rare efforts have been made for tissue section images [6] due to their complex structural morphology. Leukocytes consist of nucleus (bluish color when stained with hematoxylin and eosin (H&E) staining) within cytoplasm (pinkish color). Leukocyte can be identified with the structural and textual information of its nucleus and cytoplasm. Therefore, nuclei and cytoplasm segmentation have vital importance in leukocytes iden- tification system. Automation of leukocytes segmentation faces a set of challenges such as shape and size variability and presence of different artifacts [5,6], etc. Further, the tissue images stained with H&E staining make the segmentation process more complex. There- fore, considering the above limitations, a novel leukocytes segmen- tation strategy for images of inflamed mice skin section stained with H&E staining and acquired at 40  magnification using differential evolution (DE) algorithm is proposed. DE [12] is an evolutionary algorithm (EA) and has a vital difference from other evolutionary techniques like genetic Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/swevo Swarm and Evolutionary Computation 2210-6502/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.swevo.2013.02.003 n Corresponding author. Tel.: þ91 8989474947. E-mail addresses: saraswatmukesh@gmail.com, saraswatmukesh@iiitm.ac.in (M. Saraswat), kvarya@gmail.com (K.V. Arya), harish.sharma0107@gmail.com (H. Sharma). Swarm and Evolutionary Computation 11 (2013) 46–54