Title: Optimal hybrid classifier with fine-tuned hyper parameter and improved fuzzy C means segmentation: skin cancer detection

Authors: Sreedhar Burada; Manjunathswamy Byranahalli Eraiah; M. Sunil Kumar

Addresses: Department of Computer Science and Engineering, Don Bosco Institute of Technology, Bengaluru, India; Visvesvaraya Technological University, Belagavi, Karnataka, 560074, India ' Department of Computer Science and Engineering, Don Bosco Institute of Technology, Bengaluru, India; Visvesvaraya Technological University, Belagavi, Karnataka, 560074, India ' Department of Computer Science and Engineering, School of Computing, Mohan Babu University Formerly; Sree Vidyanikethan Engineering College, Tirupati, Andhra Pradesh 517102, India

Abstract: Skin cancer is regarded as the hazardous as well as widespread disease. Worldwide, there is been a 53% increment in present melanoma cases annually, and the mortality rate is also expected to be increasing in the coming decade. Hence, it is an urgent requirement to design a new early-detection model so that skin cancer can be more treatable without many complications. This work focuses on recognising skin cancer. The model includes the median filter (MF)-based pre-processing. The pre-processed image is subjected to a modified fuzzy C means (FCM)-based segmentation process. Finally, the recognition is done by employing a hybrid model with bi-LSTM and ANN. The proposed model's error rate was 0.091694, whereas the greatest error values for the other approaches were 0.20377 for BOA, 0.62192 for BRO, 0.170028 for ALO, 0.17168 for AOA, and 0.187915 for FIREFLY.

Keywords: skin cancer; modified FCM; bi-LSTM; improved CS-LDP; BRC-BOM algorithm.

DOI: 10.1504/IJAHUC.2024.136151

International Journal of Ad Hoc and Ubiquitous Computing, 2024 Vol.45 No.1, pp.52 - 64

Received: 26 Dec 2022
Accepted: 11 Sep 2023
Published online: 18 Jan 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article