Title: Quantitative characterisation of Plasmodium vivax in infected erythrocytes: a textural approach
Authors: Madhumala Ghosh; Devkumar Das; Chandan Chakraborty; Ajoy Kumar Ray
Addresses: School of Medical Science and Technology, Indian Institute of Technology, Kharagpur – 721302, West Bengal, India ' School of Medical Science and Technology, Indian Institute of Technology, Kharagpur – 721302, West Bengal, India ' School of Medical Science and Technology, Indian Institute of Technology, Kharagpur – 721302, West Bengal, India ' Electronics and Electrical Communication Engineering Department, Indian Institute of Technology, Kharagpur – 721302, West Bengal, India; Bengal Engineering and Science University, Botanic Garden, Howrah – 711103, West Bengal, India
Abstract: This paper aims at introducing a textural pattern analysis approach to Plasmodium vivax (P. vivax) detection from Leishman stained thin blood film. This scheme follows retrospective study design protocol where patients were selected at random in the clinic. The scheme consists of four stages - artefacts reduction, fuzzy divergence-based segmentation of P. vivax infected region(s) and normal erythrocytes, textural feature extraction using grey level co-occurrence matrix and fractal dimension, finally classification. Here, we have extracted seven features, out of which five are statistically significant in discriminating textures between malaria and normal classes based on light microscopic blood images at 100× resolutions. Finally, Bayesian and support vector machine-based classifiers are trained and validated with 100 cases and 100 control subjects. In effect, it is hereby observed that the significant textural features lead to discriminate P. vivax with 95% and 98% accuracies for SVM and Bayesian classifiers respectively. Results are studied and compared.
Keywords: parasitaemia; infected erythrocytes; Plasmodium vivax; P. vivax detection; fuzzy divergence; SVM; support vector machines; Bayesian classification; sensitivity; specificity; textural pattern analysis; thin blood film; artefacts reduction; feature extraction; malaria; microscopic blood images; textural features.
International Journal of Artificial Intelligence and Soft Computing, 2013 Vol.3 No.3, pp.203 - 221
Available online: 18 Apr 2013 *Full-text access for editors Access for subscribers Purchase this article Comment on this article