Epiderma Shield: CNN-Based Classification of Skin Diseases

Abstract

Skin diseases are a medical condition that affects the skin. The cause of these diseases might be a virus, infection, bacteria, allergy, etc. These diseases cause itchiness, rashes, inflammation, or skin changes. To recognize and differentiate these skin conditions, we have some traditional methods like blood testing and skin scraping and some modern technology too, like lasers and microscopes, but the problem is either they are less accurate or very expensive. To overcome these, we can introduce AI in the health sector, which makes tasks very easy and cheap. The introduction of a deep learning model integrated with image preprocessing in computer vision provides remote access ith the ability to detect various skin-related diseases at early stages.

Introduction

Skin diseases are common, especially in tropical and subtropical regions, due to high temperature and humidity. Negligence and lack of awareness—especially among men and rural populations—lead to severe conditions like skin cancer. A shortage of dermatologists and low public knowledge further complicate timely diagnosis. Technological solutions, such as AI-based image analysis, offer a promising way to overcome these issues.

This study proposes a CNN-based model to automatically detect two categories of skin-related diseases from images:

1. Alopecia and other hair diseases (caused by aging, heredity, hormonal changes)

2. Herpes, HPV, and other sexually transmitted diseases (STDs) (caused by bacteria, viruses, parasites)

Using a dataset from Kaggle (“Dermnet”), the model was trained on 80% of 644 images and validated/tested on the rest. Preprocessing steps included resizing, normalization, augmentation, and segmentation to improve model performance.

The CNN architecture consists of 5 convolutional layers (with 32 and 64 filters), max-pooling layers to reduce dimensionality, and fully connected layers for final classification. This model helps simplify disease detection by allowing users to upload images and receive predictions, making diagnosis more accessible.

The literature review highlights similar approaches using machine learning and deep learning (especially CNNs), their applications in skin disease detection, and the importance of image preprocessing and high-quality datasets.

Conclusion

Convolutional Neural Networks (CNNs) were used in the study\'s deep learning-based model to identify two types of skin diseases, designated A and B. With a testing accuracy of 98.44%, the researchers came to the conclusion that CNN was the best method for this classification task. According to the suggestion, the healthcare industry could be significantly impacted by such a system, especially since it would allow patients to receive quick disease detection with little effort[18]. The authors also pointed out that early diagnosis and treatment are still difficult in rural areas of India, where literacy rates are relatively low . They thought that in these underprivileged areas, this model might help with better early skin disease detection and treatment[19].

References

[1] S. B. Patil and S. K. Bodhe, “Skin disease detection using image processing and machine learning,” International Conference on Computing Communication Control and Automation (ICCUBEA), Pune, India, 2015, pp. 1–6, doi: 10.1109/ICCUBEA.2015.153. [2] M. G. Shin, S. R. Kim, and J. Y. Lee, “Diagnosis of hair loss using convolutional neural network,” in *Proc. IEEE Int. Conf. Consumer Electronics (ICCE)*, Las Vegas, NV, USA, 2020, pp. 1–2, doi: 10.1109/ICCE46568.2020.9043002. [3] A. L. Taru, A. Jain, and A. Sharma, “Automatic detection of sexually transmitted infections using deep learning,” in *Proc. IEEE Int. Conf. Computational Intelligence and Communication Technology (CICT)*, Ghaziabad, India, 2019, pp. 1–6, doi: 10.1109/CICT48419.2019.9066175. [4] S. Mahbod, G. Schaefer, R. Ecker, and A. Ellinger, “Skin disease classification using hybrid deep neural networks,” in *Proc. IEEE Int. Conf. Image Processing (ICIP)*, Taipei, Taiwan, 2019, pp. 2891–2895, doi: 10.1109/ICIP.2019.8803542. [5] K. S. Rao, P. Sharma, and A. Mehta, “Skin disease detection using CNN for medical image analysis,” in *Proc. IEEE Int. Conf. Computational Intelligence and Communication Technology (CICT)*, Ghaziabad, India, 2020, pp. 1–6, doi: 10.1109/CICT48419.2020.9066187. [6] M. K. Alam, M. M. Rahman, and M. A. H. Akhand, “Eczema detection and severity prediction using image processing and machine learning techniques,” in *Proc. IEEE Int. Conf. Imaging, Vision & Pattern Recognition (icIVPR)*, Dhaka, Bangladesh, 2019, pp. 1–6, doi: 10.1109/icIVPR48127.2019.9037592. [7] E. Nasr-Esfahani, S. Samavi, N. Karimi, R. Soroushmehr, S. Ward, and K. Najarian, “Melanoma detection by analysis of clinical images using convolutional neural network,” in *Proc. IEEE Int. Conf. Engineering in Medicine and Biology Society (EMBC)*, Orlando, FL, USA, 2016, pp. 1373–1376, doi: 10.1109/EMBC.2016.7591002. [8] A. Kumar and M. Singh, “Skin cancer detection using neural networks and support vector machine,” in *Proc. IEEE Int. Conf. Computing, Communication and Automation (ICCCA)*, Greater Noida, India, 2016, pp. 1245–1250, doi: 10.1109/CCAA.2016.7813912. [9] A. R. S. Amit, P. Sharma, and S. K. Gupta, “Potato leaf disease detection using CNN-based transfer learning models,” in *Proc. IEEE Int. Conf. Recent Advances in Intelligent Computational Systems (RAICS)*, Trivandrum, India, Jul. 2022, pp. 187–192, doi: 10.1109/RAICS54872.2022.9885283. [10] S. Goel, “Dermnet: Skin disease image dataset,” Kaggle, 2020. [Online]. Available: https://www.kaggle.com/datasets/shubhamgoel27/dermnet [11] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet classification with deep convolutional neural networks,” in *Proc. Advances in Neural Information Processing Systems (NeurIPS)*, Lake Tahoe, NV, USA, 2012, pp. 1097–1105. [12] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” *Nature*, vol. 521, no. 7553, pp. 436–444, May 2015, doi: 10.1038/nature14539. [13] D. L. Sculley, “A Comprehensive Guide to Convolutional Neural Networks — The ELI5 Way,” Saturn Cloud, Sep. 22, 2023. [Online]. Available: https://saturncloud.io/blog/a-comprehensive-guide-to-convolutional-neural-networks-the-eli5-way/ [14] R. Aditi and S. Kaur, “Skin disease classification using SVM,” in *Proc. IEEE Int. Conf. on Computational Intelligence*, New Delhi, India, 2021, pp. 123–128. [15] P. Gaikwad and V. Musande, “Skin disease detection using convolutional neural networks,” in *Proc. IEEE Int. Conf. on Emerging Trends in Computer Science and Engineering*, Pune, India, 2023, pp. 45–50. [16] S. Bangal, R. Patil, and M. Joshi, “Automated skin disease classification using CNN,” in *Proc. IEEE Int. Conf. on Machine Learning and Applications*, Mumbai, India, 2022, pp. 210–215. [17] M. Tan and Q. Le, “EfficientNet: Rethinking model scaling for convolutional neural networks,” in *Proc. 36th Int. Conf. Machine Learning (ICML)*, Long Beach, CA, USA, 2019, pp. 6105–6114. [18] S. J. Smith and R. Kumar, “Impact of AI-based diagnostic systems on healthcare delivery,” in *Proc. IEEE Int. Conf. Healthcare Informatics*, Chicago, IL, USA, 2021, pp. 145–150. [19] R. Singh, A. Sharma, and P. Gupta, “AI-based skin disease diagnosis: Challenges and opportunities in rural India,” in *Proc. IEEE Int. Conf. on Health Informatics*, Bengaluru, India, 2022, pp. 87–92.

Copyright

Copyright © 2025 Adarsh Gupta, Aditya Kukreti, Sanchi Kaushik. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.