Authors:
V. Asha, K. Aravinda, M. Arunadevi Thirumalraj, S. Gopikha, Amarilys González García

Addresses:
Department of Computer Application, New Horizon College of Engineering, Bengaluru, Karnataka, India. Department of Electronics and Communication Engineering, New Horizon College of Engineering, Bengaluru, Karnataka, India. Department of Computer Science and Engineering, Karunya Institute of Technology and Science, Coimbatore, Tamil Nadu, India. Department of Computer Science and Business Management, Saranathan College of Engineering, Tiruchirappalli, Tamil Nadu, India. Department of Information Technology, St. Joseph's College of Engineering, Chennai, Tamil Nadu, India. Department of Research and Development, Placental Histotherapy Center, Havana, Cuba.

Abstract:

Millions of individuals worldwide have deadly skin cancer. Prognosis and treatment improve with early identification and correct diagnosis. CNNs have improved medical image processing. This study presents a CNN-based method for skin cancer detection. This study employed two publicly available benchmark datasets: ISIC 2020 and HAM10000. The recommended method involves feature classification, segmentation, picture preprocessing, and hyperparameter tweaking for classification. Image brightness data was used to create a unique contrast enhancement method. UNet architecture is segmented next. The feature classification model used the Tok-MLP with dynamic sparse attention (MFETok-MLP) and Multi-scale Feature Extraction modules. This classification improved the model's ACC and training resilience. The Fast Recurrent Neural Network classified lesion images. The MPOA fixed the FRNN classifier hyperparameters to develop an excellent skin cancer prediction ACC. The suggested skin cancer detection model achieved excellent ACCs of 99.78% and 99.81% on the ISIC-2020 and HAM10000 datasets, respectively. These data demonstrate that the proposed method is more effective than current methods, helping dermatologists and other medical professionals detect skin cancer.

Keywords: Skin Cancer; Contrast Enhancement; Tokenised Multilayer Perceptron (Tok-MLP); Fast Recurrent Neural Network; Marine Predators' Optimisation Algorithm (MPOA); Convolutional neural networks (CNNs).

Received on: 14/12/2024, Revised on: 09/03/2025, Accepted on: 24/04/2025, Published on: 09/12/2025

DOI: 10.69888/FTSCS.2025.000524

FMDB Transactions on Sustainable Computing Systems, 2025 Vol. 3 No. 4, Pages: 215-232

• Views : 54
• Downloads : 7

Download PDF