Schistosomal Hepatic Fibrosis Classification

Schistosomal Hepatic Fibrosis Classification

Dalia S. Ashour, Dina M. Abou Rayia, Nilanjan Dey, Amira S. Ashour, Ahmed Refaat Hawas, Manar B. Alotaibi
Copyright: © 2018 |Pages: 17
DOI: 10.4018/IJNCR.2018040101
(Individual Articles)
No Current Special Offers


Schistosomiasis is serious liver tissues' parasitic disease that leads to liver fibrosis. Microscopic liver tissue images at different stages can be used for assessment of the fibrosis level. In the current article, the different stages of granuloma were classified after features extraction. Statistical features extraction was used to extract the significant features that characterized each stage. Afterward, different classifiers, namely the Decision Tree, Nearest Neighbor and the Neural Network are employed to carry out the classification process. The results established that the cubic k-NN, cosine k-NN and medium k-NN classifiers achieved superior classification accuracy compared to the other classifiers with 88.3% accuracy value.
Article Preview


Schistosomiasis is a serious endemic disease in 78 countries worldwide especially in tropical and subtropical areas. It affects more than 200 million people and about 800 million populations are at a great risk (Yorston & McGavin, 2009; WHO, 2014). Liver fibrosis quantitative assessment via image analysis was implemented for enhancing conventional assessment (Gailhouste et al., 2010; Tai et al., 2009). The quantitative assessment advantages are the gain of minimization of inter-observation variations and early detection of liver fibrosis stage (Bedossa, 2010).

Stanciu et al. (2014) used nonlinear microscopy for liver fibrosis quantitative assessment. In the microscopic images, the cellular and tissue information quantification of fibrosis progression have been used for diagnostic purposes. However, new diagnostic techniques should be established to overcome the limited resources (Mabey et al., 2004).

Recently, ambient intelligence (AmI) is an evolving restraint that passes intelligence to our daily life environments to make these environments sensitive to the humans. Several researches are interested to develop different systems based on AmI and image processing (Samanta et al., 2016; Saba et al., 2016; Belgharb, & Boufaida, 2017; Juneja et al., 2017; Alenljung et al., 2017; Acharjya, & Anitha, 2017; Sharma, & Virmani, 2017; Manogaran, & Lopez, 2017; Ahmed et al., 2017). AmI research is based on developments in sensors, prevalent computing, and artificial intelligence. Artificial intelligence techniques along with medical image processing are employed for computer-aided diagnosis for the liver tissues classification. Ghoneim (2011) investigated the texture analysis (TA) accuracy results at different resolutions on three color spaces, namely RGB (Red, Green, Blue), conventional grey scale and the Hue-Saturation-Intensity (HSI) for texture classification of liver images. The RGB provided accurate results at low resolution. Additionally, the green channel provided several characterizing features for the fibrosis images. At high resolution, the grey scale space formed superior results; though, errors increased with decreased resolution. However, the HSI space provided high error percentage at all resolutions. Thus, it is insufficient for fibrosis characterization.

Therefore, the contribution of current work is to use the conventional microscopy images for assessment and staging of liver fibrosis using image processing techniques based on statistical features extraction and classification using different classifiers, namely Decision Tree (DT), Nearest Neighbor (k-NN) and the Neural Network (NN). Such classification is employed to quantify and automate the liver fibrosis staging using animal model liver samples. The comparative study of the decision tree, neural network and the nearest neighbor classifiers performance of liver fibrosis staging are performed.

Complete Article List

Search this Journal:
Volume 12: 1 Issue (2023): Forthcoming, Available for Pre-Order
Volume 11: 4 Issues (2022): 1 Released, 3 Forthcoming
Volume 10: 4 Issues (2021)
Volume 9: 4 Issues (2020)
Volume 8: 4 Issues (2019)
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