Harnessing Deep Learning for Scalp and Hair Disease Classification: A Comparative Study of Convolutional Neural Networks Architectures

Dang Nguyen Nam Anh; Pham Ngoc Gia; Binh Nguyen Le Nguyen; An Mai; Nguyen Thi Minh Hien; Nguyen Tan Viet Tuyen

Dang Nguyen Nam Anh School of Computer Science and Engineering International University, Ho Chi Minh City, Vietnam ,Vietnam National University Ho Chi Minh City, Vietnam
Pham Ngoc Gia Tien Giang University, Tien Giang Province, Vietnam ,Tra Vinh University, Tra Vinh Province, Vietnam
Binh Nguyen Le Nguyen School of Computer Science and Engineering International University, Ho Chi Minh City, Vietnam ,Vietnam National University Ho Chi Minh City, Vietnam
An Mai School of Computer Science and Engineering International University, Ho Chi Minh City, Vietnam ,Vietnam National University Ho Chi Minh City, Vietnam
Nguyen Thi Minh Hien FPT University, Ho Chi Minh City, Vietnam
Nguyen Tan Viet Tuyen School of Electronics and Computer Science, University of Southampton, Southampton United Kingdom

Abstract

Scalp and hair diseases, affecting millions worldwide, pose significant challenges regarding accurate diagnosis and effective treatment. Traditionally reliant on expert evaluation, these conditions can often be misdiagnosed due to their complex and overlapping symptoms. In recent times, especially in information technology, convolutional neural networks (CNNS) have become more prominent thanks to their ability to analyse and process image data for classification and recognition tasks. CNNs learn to recognize patterns from images through convolutional layers to detect characteristic features in image sand have revolutionized the field of image recognition, offering promising applications in medical diagnostics. Despite their potential, few studies have thoroughly explored the capabilities of multiple CNN architecturesin the context of dermatology. This study aims to bridge this gap by evaluating the effectiveness of several CNN models—VGG16, VGG19 , Inception-V3, ResNet50, and ResNet152—in detecting scalp and hair diseases. The findings indicate that VGG16 and VGG19 consistently outper for mother models in accuracy across all disease categories, demon strating their robustness and reliability for this application. By providing a comparative analysis of these architectures with a user interface (UI), we seek to advance automated diagnostic methods, ultimately enhancing clinical decision-making and patient care.

References

[1] Takagi, Masaya., et al. (2023). Invention of automated numerical algorithm adopting binarization for the evaluation of scalp hair coverage: An image analysis providing a substitute for phototrichogram and global photography assessment for hair diseases. Journal of Dermatological Science, 112(2), 92–98. [2] Wortmann, F. J., et al. (2013). Considerations of a consistent model for the reflection of light from human hair. [3] Hordinsky, Maria K. (2013). Overview of Alopecia Areata. Journal of Investigative Dermatology Symposium Proceedings, 16(1). https://doi.org/10.1038/JIDSYMP.2013.4 [4] Chieh, Chen. (2023). Alopecia areata. Journal of Vaccines and Immunology, 9(1), 001–005. https://doi.org/ 10.17352/jvi.000054 [5] Kormeili, T., Lowe, N., Yamauchi, P. (2004). Psoriasis: Immunopathogenesis and evolving immunomodulators and systemic therapies; U.S. experiences. British Journal of Dermatology, 151. https://doi.org/10.1111/j.13652133.2004.06009.x [6] Sampaio, Ana Luisa Sobral Bittencourt., et al. (2011). Seborrheic dermatitis. Anais Brasileiros de Dermatologia, 86, 1061–1074. [7] Borgers, M., Degreef, H., Cauwenbergh, G. (2005). Fungal infections of the skin: Infection process and antimycotic therapy. Current Drug Targets, 6(8) 849–862. [8] Mirzoyev, Sultan A., et al. (2013). Lifetime incidence risk of Alopecia Areata estimated at 2.1 percent by Rochester Epidemiology Project, 1990–2009. The Journal of Investigative Dermatology, 134, 1141–1142. https:/ /doi.org/10.1038/jid.2013.464[9] Wolff, H., Fischer, T. W., Blume-Peytavi, U. (2016). The diagnosis and treatment of hair and scalp diseases. DtschArztebl Int, 113(21) 377–386. [10] Wang, W.-C., Chen, L.-B., Chang, W.-J. (2018). Development and experimental evaluation of machinelearning techniques for an intelligent hairy scalp detection system. Applied Sciences, 8, 853. [11] Rao, G. Madhukar., et al. (2023). A hybrid deep learning strategy for image-based automated prognosis of skin disease. Soft Computing, 1–12. [12] Ahammed, Mostafiz., Al Mamun, Md., Uddin, Mohammad Shorif. (2022). A machine learning approach for skin disease detection and classification using image segmentation. Healthcare Analytics, 2, 100122. [13] Tschandl, Philipp., Rosendahl, Cliff., Kittler, Harald. (2018). The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Scientific Data, 5 (1) 1–9. [14] Kim, M., Gil, Y., Kim, Y., Kim, J. (2023). Deep-learning-based scalp image analysis using limited data. Electronics, 12, 1380. [15] Roy, Mrinmoy., Protity, Anica Tasnim. (2023). Hair and scalp disease detection using machine learning and image processing. EJ-Compute, 3(1) 7–13. [16] Busireddy Seshakagari Haranadha Reddy. (2025). Deep learning-based detection of hair and scalp diseases using CNN and image processing. Milestone Transactions on Medical Technometrics, 3(1), 145–155. [17] Sultanpure, Kavita., Shirsath, Bhairavi., Bhande, Bhakti., Sawai, Harshada., Gawade, Srushti., Samgir, Suraj. (2024). Hair and scalp disease detection using deep learning. arXiv preprint. https://arxiv.org/abs/ 2403.07940 [18] Tiwari, Hrishabh., Moolchandani, Jatin., Mantri, Shamla. (2023). Diagnosis of scalp disorders using machine learning and deep learning approach: A review. arXiv preprint. https://arxiv.org/abs/2308.07052 [19] Baresary, D., Jambu, M., Bansal, A., Singh Brar, T. P., Kaushik, P., Mehta, S. (2024). Hair disease detection with deep learning: A VGG16-driven model for precise classification of hair and scalp conditions. In 2024 International Conference on Decision Aid Sciences and Applications (DASA) (pp. 1–6). IEEE. [20] Rafay, Abdul., Hussain, Waqar. (2023). EfficientSkinDis: An EfficientNet-based classification model for a large manually curated dataset of 31 skin diseases. Biomedical Signal Processing and Control, 85, 104869. [21] Chang, W.-J., Chen, L.-B., Chen, M.-C., Chiu, Y.-C., Lin, J.-Y. (2020). ScalpEye: A deep learning-based scalp hair inspection and diagnosis system for scalp health. IEEE Access, 8, 134826–134837. [22] Annamalai, Sundar. Hair disease dataset. Retrieved from https://www.kaggle.com/datasets/ sundarannamalai/hair-diseases/data[23] Ali, Luqman., et al. (2021). Performance evaluation of deep CNN-based crack detection and localization techniques for concrete structures. Sensors, 21(5) 1688. [24] Jhong, Sin-Ye., Yang, Po-Yen., Hsia, Chih-Hsien. (2022). An expert smart scalp inspection system using deep learning. Sensors and Materials, 34(4) 1265–1274. [25] Chowdhury, Mohammad Sayem., Sultan, Tofayet., Jahan, Nusrat., Mridha, Muhammad Firoz., Safran, Mejdl., Alfarhood, Sultan., Che, Dunren. (2024). Leveraging deep neural networks to uncover unprecedented levels of precision in the diagnosis of hair and scalp disorders. Skin Research and Technology, 30(4), e13660. [26] Khan, V., Subramaniam, K. (2023). An effective optimization-based convolutional neural network for effective hair and scalp detection system. In 2023 1st International Conference on Cognitive Computing and Engineering Education (ICCCEE) (pp. 1–8).