Open Access

ARTICLE

An Efficient Hybrid Optimization for Skin Cancer Detection Using PNN Classifier

J. Jaculin Femil^1,*, T. Jaya²

1 Ponjesly College of Engineering, Nagercoil, Tamilnadu, India
2 CSI Institute of Technology Thovalai, India

* Corresponding Author: J. Jaculin Femil. Email:

Computer Systems Science and Engineering 2023, 45(3), 2919-2934. https://doi.org/10.32604/csse.2023.032935

Received 02 June 2022; Accepted 12 July 2022; Issue published 21 December 2022

Abstract

The necessity of on-time cancer detection is extremely high in the recent days as it becomes a threat to human life. The skin cancer is considered as one of the dangerous diseases among other types of cancer since it causes severe health impacts on human beings and hence it is highly mandatory to detect the skin cancer in the early stage for providing adequate treatment. Therefore, an effective image processing approach is employed in this present study for the accurate detection of skin cancer. Initially, the dermoscopy images of skin lesions are retrieved and processed by eliminating the noises with the assistance of Gabor filter. Then, the pre-processed dermoscopy image is segmented into multiple regions by implementing cascaded Fuzzy C-Means (FCM) algorithm, which involves in improving the reliability of cancer detection. The A Gabor Response Co-occurrence Matrix (GRCM) is used to extract melanoma parameters in an efficient manner. A hybrid Particle Swarm Optimization (PSO)-Whale Optimization is then utilized for efficiently optimizing the extracted features. Finally, the features are significantly classified with the assistance of Probabilistic Neural Network (PNN) classifier for classifying the stages of skin lesion in an optimal manner. The whole work is stimulated in MATLAB and the attained outcomes have proved that the introduced approach delivers optimal results with maximal accuracy of 97.83%.

---

Keywords

---