Segmentation of Peripheral Blood Smear Images Using Tissue-Like P Systems

Segmentation of Peripheral Blood Smear Images Using Tissue-Like P Systems

Feminna Sheeba (Madras Christian College, India), Atulya K. Nagar (Liverpool Hope University, UK), Robinson Thamburaj (Madras Christian College, India) and Joy John Mammen (Christian Medical College, India)
Copyright: © 2012 |Pages: 12
DOI: 10.4018/jncr.2012010102


The tissue-like P Systems, which are based on the methodology of cell and tissue behavior in a human body, are used in various areas of computation. Segmentation of medical images is one such area where these systems can be used to identify various details and objects in those images. It is a highly challenging process, especially when dealing with blood smear images, which have a very complex cell structure. In order to analyze each object individually and to avoid the cumbersome and error-prone existing manual methods, images can be segmented using appropriate automated segmentation techniques. The proposed work aims at segmenting the nuclei of the White Blood Cells (WBCs) of the peripheral blood smear images, using tissue-like P Systems, which can help to identify various pathological conditions. In the first approach, segmentation is color based. The second approach is intensity based. In the third approach, morphology is used to strengthen the findings from the results.
Article Preview


Segmentation in medical images is a crucial technique used in image analysis and is very helpful in the diagnosis of various diseases. The cells in peripheral blood smear images are quite complex in nature and therefore need advanced segmentation techniques in order to identify and analyze the individual objects in the image. A simple segmentation technique is used to separate WBCs from other objects and the background of the image using thresholding and regional context information (Sheeba, Hannah, & Mammen, 2010; Liao & Deng, 2002). Another simple method is to use morphological analysis (Sheeba, Hannah, & Mammen, 2010; Sadeghian et al., 2009; Angulo & Flandrin, 2003). As the nuclei in WBCs are darker than all the objects and the background of the image, segmentation of the nuclei was done by selecting them with a particular threshold value.Thresholding cannot be used for segmenting overlapping objects. An important segmentation technique that is used to segment such overlapping objects is the watershed transform (Sheeba, Thamburaj, Mammen, Hannah, & Nagar, 2011; Dorini & Sahoolizadeh, 2009; Minetto & Leite, 2007).Overlapping objects can also be segmented by finding interesting points in the objects using chain code algorithms (Nithya &Nirmala, 2012). Isolating WBCs from other objects and background can be done based on the colorby applying various color segmentation methods (Hiremath, Bannigidad, & Geeta, 2010; Mohapatra & Patra, 2010). This is also a very simple technique, provided the WBCs and Red Blood Cells (RBCs) are not almost of the same color. Region growing is another segmentation method widely used. Region growing and merging can be used to find the contours of objects in the images (Banerjee, Bhattacharjee, & Chowdhur, 2010). Based on region growing, the homogeneity criterion is achieved by using the gray value and standard deviation of the regions (Pohle &Toennies, 2001). Certain segmentation methods used the snake algorithm (Ongun, Halici, Leblebicioglu, Atalay, Beksac, &Beksac, 2002). Finding the centre of a mass as the initial position for the algorithm to progress is the challenge in such algorithms. Border finding is done using graphs, when cells touch one another (Danek et al., 2009; Ritter & Cooper, 2002). Teager energy based segmentation resulted in better segmentation of nucleus (Ravikumar, Joseph, &Sreenivas, 2002). Fuzzy logic can also be applied to segment WBCs (Mohapatra, Patra & Satpathi, 2010; Chinwarapath, Sanpanich, Pintavirooj, Sangworasil, &Tosranon, 2008; Theera-Umpon, 2005).

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 8: 4 Issues (2019): 2 Released, 2 Forthcoming
Volume 7: 4 Issues (2018)
Volume 6: 2 Issues (2017)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing