Skin Cancer Detection Using a CNN Model DermaNet with a Comparative Analysis of Activation Functions, Optimizers and Data Balancing Techniques

International Journal of Electronics and Communication Engineering

© 2025 by SSRG - IJECE Journal

Volume 12 Issue 6

Year of Publication : 2025

Authors : B. Lakshmi Prasanna, Ravi Boda, R. Murali Prasad

10.14445/23488549/IJECE-V12I6P110

How to Cite?

B. Lakshmi Prasanna, Ravi Boda, R. Murali Prasad, "Skin Cancer Detection Using a CNN Model DermaNet with a Comparative Analysis of Activation Functions, Optimizers and Data Balancing Techniques," SSRG International Journal of Electronics and Communication Engineering, vol. 12,  no. 6, pp. 119-131, 2025. Crossref, https://doi.org/10.14445/23488549/IJECE-V12I6P110

Abstract:

A serious worldwide health issue is skin cancer, which necessitates prompt and precise diagnosis. This research uses the HAM10000 dataset to present Derma Net, a customized convolutional neural network for classifying skin cancer. DermaNet was tested using balanced and unbalanced datasets, various activation functions, and multiple optimizers to identify the optimal configuration. Poor performance in minority classes resulted from initial training on the unbalanced dataset. Classification significantly improved after oversampling. After testing various optimizers such as Adam, Adamax, Adagrad, Nadam, and RMSprop, as well as activation functions such as ReLU, Clipped ReLU, Hyperbolic Tangent, Leaky ReLU, ELU, and PReLU, the combination of ELU and Nadam was found to yield the best results, with 97.0% accuracy, 92.6% precision, 94.9% recall,93.8% F1-score and 0.98 AUC. This combination offered excellent precision for medical applications by lowering false positives and high sensitivity by limiting false negatives. Our results highlight the importance of optimizer adjusting, activation selection, and dataset balancing to diagnose skin cancer. Using ELU and Nadam, Derma Net is a promising AI-based diagnostic tool that can help dermatologists identify issues early and enhance patient outcomes.

Keywords:

Dataset balancing, Derma net, Deep Learning, Activation functions and optimizers, Skin cancer classification.

References:

[1] Yali Nie et al., “Recent Advances in Diagnosis of Skin Lesions Using Dermoscopic Images Based on Deep Learning,” IEEE Access, vol. 10, pp. 95716-95747, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[2] Muhammad Qasim Khan et al., “Classification of Melanoma and Nevus in Digital Images for Diagnosis of Skin Cancer,” IEEE Access, vol. 7, pp. 90132-90144, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[3] Skin Cancer Statistics Melanoma of the Skin is the 19th Most Common Cancer Worldwide, World Cancer Research Fund, 2020. [Online]. Available: https://www.wcrf.org/dietandcancer/cancer-trends/skin-cancer-statistics
[4] G. Alden Holmes et al., “Using Dermoscopy to Identify Melanoma 1200 and Improve Diagnostic Discrimination,” Federal Practitioner, vol. 35, no. 4, pp. 39-45, 2018.
[Google Scholar] [Publisher Link]
[5] Duggani Keerthana et al., “Hybrid Convolutional Neural Networks with SVM Classifier for Classification of Skin Cancer,” Biomedical Engineering Advances, vol. 5, pp. 1-8, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[6] N. Priyadharshini et al., “A Novel Hybrid Extreme Learning Machine and Teaching–Learning-Based Optimization Algorithm for Skin Cancer Detection,” Healthcare Analytics, vol. 3, pp. 1-10, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[7] Van-Dung Hoang, Xuan-Thuy Vo, and Kang-Hyun Jo, “Categorical Weighting Domination for Imbalanced Classification with Skin Cancer in Intelligent Healthcare Systems,” IEEE Access, vol. 11, pp. 105170-105181, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[8] Skin Cancer MNIST: HAM10000, Kaggle. [Online]. Available. https://www.kaggle.com/datasets/kmader/skin-cancer-mnist-ham10000
[9] Stephanie S. Noronha et al., “Deep Learning-Based Dermatological Condition Detection: A Systematic Review with Recent Methods, Datasets, Challenges, and Future Directions,” IEEE Access, vol. 11, pp. 140348-140381, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[10] C.H. Qiu et al., “Application Review of Artificial Intelligence in Medical Images Aided Diagnosis,” Space Medicine and Medical Engineering, vol. 34, no. 5 pp. 407-414, 2021.
[Google Scholar]
[11] Li Zhang et al., “A Deep Learning Outline Aimed at Prompt Skin Cancer Detection Utilizing Gated Recurrent Unit Networks and Improved Orca Predation Algorithm,” Biomedical Signal Processing and Control, vol. 90, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[12] Ashwani Kumar et al., “A Novel Skin Cancer Detection Model using Modified Finch Deep CNN Classifier Model,” Scientific Reports, vol. 14, pp. 1-19, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[13] Krishna Mridha et al., “An Interpretable Skin Cancer Classification Using Optimized Convolutional Neural Network for a Smart Healthcare System,” IEEE Access, vol. 11, pp. 41003-41018, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[14] H.L. Gururaj et al., “DeepSkin: A Deep Learning Approach for Skin Cancer Classification,” IEEE Access, vol. 11, pp. 50205-50214, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[15] E. Gomathi et al., “Skin Cancer Detection Using Dual Optimization based Deep Learning Network,” Biomedical Signal Processing and Control, vol. 84, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[16] S.P. Godlin Jasil, and V. Ulagamuthalvi, “A Hybrid CNN Architecture for Skin Lesion Classification using Deep Learning,” Soft Computing, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[17] Sachin Gupta et al., “Ensemble Optimization Algorithm for the Prediction of Melanoma Skin Cancer,” Measurement: Sensors, vol. 29, pp. 1-9, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[18] Yagmur Olmez et al., “Improved PSO with Visit Table and Multiple Direction Search Strategies for Skin Cancer Image Segmentation,” IEEE Access, vol. 12, pp. 840-867, 2024.
[CrossRef] [Google Scholar] [Publisher Link]
[19] Himanshu K. Gajera, Deepak Ranjan Nayak, and Mukesh A. Zaveri, “A Comprehensive Analysis of Dermoscopy Images for Melanoma Detection via Deep CNN Features,” Biomedical Signal Processing and Control, vol. 73, no. 2, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[20] V. Auxilia Osvin Nancy et al., “Comparative Study and Analysis on Skin Cancer Detection using Machine Learning and Deep Learning Algorithms,” Multimedia Tools and Applications, vol. 82, pp. 45913-45957, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[21] Dangguo Shao, Lifan Ren, and Lei Ma, “MSF-Net: A Lightweight Multi-Scale Feature Fusion Network for Skin Lesion Segmentation,” Biomedicines, vol. 11, no. 6, pp. 1-15, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[22] Pavel A. Lyakhov, Ulyana A. Lyakhova, and Diana I. Kalita, “Multimodal Analysis of Unbalanced Dermatological Data for Skin Cancer Recognition,” IEEE Access, vol. 11, pp. 131487-131507, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[23] Arslan Akram et al., “Segmentation and Classification of Skin Lesions using Hybrid Deep Learning Method in the Internet of Medical Things,” Skin Research and Technology, vol. 29, no. 11, pp. 1-14, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[24] Maria Tamoor et al., “Skin Lesion Segmentation Using an Ensemble of Different Image Processing Methods,” Diagnostics, vol. 13, no. 16, pp. 1-14, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[25] Peng Yao et al., “Single Model Deep Learning on Imbalanced Small Datasets for Skin Lesion Classification,” IEEE Transactions on Medical Imaging, vol. 41, no. 5, pp. 1242-1254, 2022.
[CrossRef] [Google Scholar] [Publisher Link]