Color and Texture Prior Based Segmentation and Analysis of Psoriatic Disease Types Using MPSO

Abstract

The Computer-aided diagnosis (CAD) tools are needed to address the human limitation in the assessment of the severity and classification of psoriasis. Dermatologists need these kinds of systems to assist to produce high diagnostic results. The main objective of this work is to develop an automatic machine that predicts the psoriatic type and estimates the disease parameters. The proposed work includes preprocessing with the morphological closing method, Segmentation, feature extraction, and Indexing. A novel Color and Texture prior based segmentation scheme CTG-SEG with extraction ability has been implemented in the high wound region. The feature extraction process has used feature descriptors such as texture: a total of 31 texture descriptors, Color: a total of 38 color descriptors and Shape: a total of 14 shape descriptors contributing to a total of 83 features. Genetic Algorithm (GA) is employed here to opt for the features that contribute most to the expected prediction. Multiclass Particle Swarm Optimization (MPSO) classifies the types of psoriasis disease using the nearly 11 selected features. Based on Psoriasis Area and Severity Index (PASI), a Severity Score (SS) measure is calculated using the psoriatic region to assess the severity of the disease. The native collected dataset is used for experimental analysis. The proposed method is promising with an average accuracy of 95.7%, a specificity of 96% of and a sensitivity of 95.6%. The comparative analysis with the existing methods shows that the proposed method has highly contributed to psoriasis region segmentation, classification, and severity grading.

KEYWORDS:

• Feature extraction
• Genetic algorithm
• Indexing
• Psoriasis
• Segmentation

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

S.V. Anandhi

S V Anandhi received her BE degree in computer science and engineering in 2006 from Noorul Islam College of Engineering, India and ME degree in computer science and engineering in 2010 from Dr Sivanthi Aditanar College of Engineering, Tiruchendur India. Presently she is working as assistant professor in computer science and engineering at Dr Sivanthi Aditanar College of Engineering, Tiruchendur, India. Her research interest includes image processing.

G. Wiselin Jiji

G Wiselin Jiji is a professor of computer science and engineering at Dr Sivanthi Aditanar College of Engineering, Tiruchendur. She has published more than 78 scientific research papers. She is a recipient of ten national and three state awards. Her long-term research focuses on computer-aided detection (CAD) and measurement (CAM) of lesions in medical images. CAD aims at discovering the fundamental perception processes of human vision in the image-based diagnosis of lesions and developing mathematical/computational models that describe them. Her area of interests are computer-aided detection and diagnosis of abnormality, using medical images and medical image analyses, such as image enhancement, segmentation, feature extraction, object detection, and pattern recognition. Email: [email protected]