A Comparative Machine and Deep Learning Framework for Multi-Class Hair Fall and Scalp Disease Classification Using Optimized Convolutional Neural Networks
DOI:
https://doi.org/10.21015/vtcs.v14i1.2308Abstract
Scalp diseases and hair fall represent a serious threat to the world population and need to be diagnosed properly and timely to prevent dermatological complications in the long run. The traditional form of diagnosis can be founded on manual clinical examination that can be subjective, time consuming and resource consuming. The current paper introduces an AI-based system to automatically classify hair fall and scalp diseases with machine and deep learning algorithms. The data of multi-class scalp disease was used to test the effectiveness of various classification models, including Logistic Regression, k-nearest-neighbors (KNN), Random Forest, Support Vector Machine (SVM), Decision Tree, Artificial Neural Network (ANN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Convolutional Neural Network (CNN). The best discriminative performance of the Logistic Regression model was achieved with AUC of 0.975, and an overall classification accuracy of 87.97% was reached using augmentation, class-weighted learning and early stopping techniques. The proposed system has the potential to deliver an efficient and scalable decision support system to diagnose scalp disease and detect hair fall in the initial stages.
References
[1] C. H. Ho, T. Sood, and P. M. Zito, “Androgenetic alopecia,” in *StatPearls [Internet]*. Treasure Island, FL, USA: StatPearls Publishing, 2024.
[2] T. Hiroyasu, K. Hayashinuma, H. Ichikawa, and N. Yagi, “Preprocessing with image denoising and histogram equalization for endoscopy image analysis using texture analysis,” in *Proc. 37th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. (EMBC)*, Milan, Italy, 2015, pp. 789–792.
[3] Y. Chen et al., “Genetic prediction of male pattern baldness based on large independent datasets,” *Eur. J. Hum. Genet.*, vol. 31, no. 3, pp. 321–328, Mar. 2023.
[4] E. Darwin, P. A. Hirt, R. Fertig, B. Doliner, G. Delcanto, and J. J. Jimenez, “Alopecia areata: review of epidemiology, clinical features, pathogenesis, and new treatment options,” *Int. J. Trichology*, vol. 10, no. 2, pp. 51–60, Mar.–Apr. 2018.
[5] M. S. Chowdhury et al., “Leveraging deep neural networks to uncover unprecedented levels of precision in the diagnosis of hair and scalp disorders,” *Skin Res. Technol.*, vol. 30, no. 4, p. e13660, Apr. 2024.
[6] C. S. Shakeel, S. J. Khan, B. Chaudhry, S. F. Aijaz, and U. Hassan, “Classification framework for healthy hairs and alopecia areata: a machine learning (ML) approach,” *Comput. Math. Methods Med.*, vol. 2021, art. no. 1102083, 2021.
[7] C. Ha, T. Go, and W. Choi, “Intelligent healthcare platform for diagnosis of scalp and hair disorders,” *Appl. Sci.*, vol. 14, no. 5, p. 1734, Feb. 2024.
[8] J. H. Kim, S. Kwon, J. Fu, and J. H. Park, “Hair follicle classification and hair loss severity estimation using Mask R-CNN,” *J. Imaging*, vol. 8, no. 10, p. 283, Oct. 2022.
[9] K. Sultanpure, B. Shirsath, B. Bhande, H. Sawai, S. Gawade, and S. Samgir, “Hair and scalp disease detection using deep learning,” *arXiv*, arXiv:2403.07940, Mar. 2024.
[10] W. J. Chang, M. C. Chen, L. B. Chen, Y. C. Chiu, C. H. Hsu, Y. K. Ou, and Q. Chen, “A mobile device-based hairy scalp diagnosis system using deep learning techniques,” in *Proc. IEEE 2nd Global Conf. Life Sci. Technol. (LifeTech)*, Sydney, NSW, Australia, 2020, pp. 145–146.
[11] M. Di Fraia, L. Tieghi, F. Magri, G. Caro, S. Michelini, G. Pellacani, and A. Rossi, “A machine learning algorithm applied to trichoscopy for androgenic alopecia staging and severity assessment,” *Dermatol. Pract. Concept.*, vol. 13, no. 3, Jul. 2023.
[12] S. A. Farooq, A. Ali, and A. Bashir, “The prediction of hairfall pattern in a person using artificial intelligence for better care and treatment,” in *Proc. 4th Int. Conf. Innov. Pract. Technol. Manage. (ICIPTM)*, Noida, India, 2024, pp. 1–6.
[13] S. K. Khokhar, A. Qamar, and Y. Mahar, “Effects of exposure time to sun on hair fall during lockdown in Covid pandemic,” *J. Gandhara Med. Dent. Sci.*, vol. 10, no. 4, pp. 64–67, Oct.–Dec. 2023.
[14] Y. Yacoob and L. S. Davis, “Detection and analysis of hair,” *IEEE Trans. Pattern Anal. Mach. Intell.*, vol. 28, no. 7, pp. 1164–1169, Jul. 2006.
[15] M. Di Fraia et al., “A machine learning algorithm applied to trichoscopy for androgenic alopecia staging,” *Dermatol. Pract. Concept.*, vol. 13, no. 3, Jul. 2023.
[16] D. S. Park, J. E. Koo, S. Y. Oh, and H. J. Choi, “Physiological profiles of hair and scalp conditions in Korea population,” *Adv. Bioeng. Biomed. Sci. Res.*, vol. 4, no. 2, pp. 63–71, 2021.
[17] S. Y. Jhong, P. Y. Yang, and C. H. Hsia, “An attention based expert inspection system for smart scalp,” in *Proc. Asia-Pacific Signal Inf. Process. Assoc. Annu. Summit Conf. (APSIPA ASC)*, Tokyo, Japan, 2021, pp. 1678–1681.
[18] X. Liu, C. H. Chen, M. Karvela, and C. Toumazou, “A DNA-based intelligent expert system for personalised skin-health recommendations,” *IEEE J. Biomed. Health Informat.*, vol. 24, no. 11, pp. 3276–3284, Nov. 2020.
[19] C. R. Kothari and S. Patil, “Trichoscopic features in female pattern hair loss: 1-year hospital-based cross-sectional study,” *Clin. Dermatol. Rev.*, vol. 8, no. 2, pp. 95–101, May–Aug. 2024.
[20] C. S. Shakeel and S. J. Khan, “Machine learning techniques as effective methods for evaluating hair and skin assessments: a systematic review,” *Proc. Inst. Mech. Eng., Part H*, vol. 238, no. 2, pp. 132–148, Feb. 2024.
[21] C. Saraswathi and B. Pushpa, “Ab-mtedeep classifier trained with AAGAN for the identification and classification of alopecia areata,” *Eng. Technol. Appl. Sci. Res.*, vol. 13, no. 3, pp. 10895–10900, Jun. 2023.
[22] A. Egger, M. Tomic-Canic, and A. Tosti, “Advances in stem cell-based therapy for hair loss,” *CellR4*, vol. 8, 2020.
[23] M. S. Chowdhury et al., “Leveraging deep neural networks to uncover unprecedented levels of precision in the diagnosis of hair and scalp disorders,” *Skin Res. Technol.*, vol. 30, no. 4, p. e13660, Apr. 2024.
[24] E. Darwin et al., “Alopecia areata: review of epidemiology, clinical features, pathogenesis, and new treatment options,” *Int. J. Trichology*, vol. 10, no. 2, pp. 51–60, 2018.
[25] W. C. Wang, L. B. Chen, and W. J. Chang, “Development and experimental evaluation of machine-learning techniques for an intelligent hairy scalp detection system,” *Appl. Sci.*, vol. 8, no. 6, p. 853, May 2018.
[26] E. K. Ross, C. Vincenzi, and A. Tosti, “Videodermoscopy in the evaluation of hair and scalp disorders,” *J. Amer. Acad. Dermatol.*, vol. 55, no. 5, pp. 799–806, Nov. 2006.
[27] R. Nagar and R. Dhudshia, “Utility of trichoscopy to diagnose early female pattern hair loss in resource-poor setting: a cross-sectional study,” *Indian J. Dermatol. Venereol. Leprol.*, vol. 85, p. 681, 2019.
[28] R. Behal, P. Priya, A. Kapoor, and A. Mishra, “Hair loss stage prediction using deep learning,” in *Proc. IEEE Int. Conf. Comput. Power Commun. Technol. (IC2PCT)*, vol. 5, Greater Noida, India, 2024, pp. 1263–1267.
[29] J. H. Kim, S. Kwon, J. Fu, and J. H. Park, “Hair follicle classification and hair loss severity estimation using Mask R-CNN,” *J. Imaging*, vol. 8, no. 10, p. 283, Oct. 2022.
[30] D. Banerjee, V. Kukreja, D. Bordoloi, and A. Choudhary, “Enhanced hair disease classification using deep learning,” in *Proc. 11th Int. Conf. Signal Process. Integr. Netw. (SPIN)*, Noida, India, 2024, pp. 58–63.
[31] L. P. Liu, M. A. Wariboko, X. Hu, Z. H. Wang, Q. Wu, and Y. M. Li, “Factors associated with early-onset androgenetic alopecia: a scoping review,” *PLoS ONE*, vol. 19, no. 3, p. e0299212, Mar. 2024.
[32] N. Jain, B. Doshi, and U. Khopkar, “Trichoscopy in alopecias: diagnosis simplified,” *Int. J. Trichology*, vol. 5, no. 4, pp. 170–178, Oct.–Dec. 2013.
[33] R. Xiao and L. N. Lee, “Updated review of treatment of androgenetic alopecia,” *Facial Plast. Surg. Clin.*, vol. 32, no. 3, pp. 375–384, Aug. 2024.
[34] N. Natarelli, N. Gahoonia, and R. K. Sivamani, “Integrative and mechanistic approach to the hair growth cycle and hair loss,” *J. Clin. Med.*, vol. 12, no. 3, p. 893, Jan. 2023.
[35] B. M. Piraccini et al., “Efficacy and safety of topical finasteride spray solution for male androgenetic alopecia: a phase III, randomized, controlled clinical trial,” *J. Eur. Acad. Dermatol. Venereol.*, vol. 36, no. 2, pp. 286–294, Feb. 2022.
[36] C. C. Chen, M. V. Plikus, P. C. Tang, R. B. Widelitz, and C. M. Chuong, “The modulatable stem cell niche: tissue interactions during hair and feather follicle regeneration,” *J. Mol. Biol.*, vol. 428, no. 7, pp. 1423–1440, Apr. 2016.
[37] P. Bouhanna, “Multifactorial classification of male and female androgenetic alopecia,” *Dermatol. Surg.*, vol. 26, no. 6, pp. 555–561, Jun. 2000.
[38] I. Al-Aradi and M. Al-Ghareeb, “Hair fall: common causes and treatment modalities,” *Res. Gate*, vol. 16, no. 1, pp. 9–15, 2005.
[39] L. Srinivasan, A. Jeevika, R. K. Navina, and S. Priyadharshini, “An enhanced stress based hairfall detection and prevention using KNN and machine learning techniques,” in *Proc. 7th Int. Conf. Trends Electron. Inform. (ICOEI)*, Tirunelveli, India, 2023, pp. 1110–1115.
[40] H. Pant, G. Joshi, J. Pant, D. Singh, S. Maurya, and H. R. Goyal, “Advancing dermatology: deep convolutional neural networks for timely hair disease diagnosis,” in *Proc. IEEE 3rd World Conf. Appl. Intell. Comput. (AIC)*, Ghaziabad, India, 2024, pp. 761–765.
[41] O. O. Ayanlowo, “Scalp and hair disorders at the dermatology outpatient clinic of a tertiary hospital,” *Port Harcourt Med. J.*, vol. 11, no. 3, pp. 127–133, Sep.–Dec. 2017.
[42] F. Canpolat, “Infections, infestations and neoplasms of the scalp,” in *Hair and Scalp Disorders*, Z. Kutlubay and S. Serdaroglu, Eds. London, U.K.: IntechOpen, 2017, ch. 12, pp. 199–219.
[43] H. Wolff, T. W. Fischer, and U. Blume-Peytavi, “The diagnosis and treatment of hair and scalp diseases,” *Dtsch. Arztebl. Int.*, vol. 113, no. 21, p. 377, May 2016.
[44] A. Rajab, H. Farman, N. Islam, D. Syed, M. A. Elmagzoub, A. Shaikh, and M. Alrizq, “Flood forecasting by using machine learning: a study leveraging historic climatic records of Bangladesh,” *Water*, vol. 15, no. 22, p. 3970, Nov. 2023.
[45] H. Farman, A. W. Khan, S. Ahmed, D. Khan, M. Imran, and P. Bajaj, “An analysis of supervised machine learning techniques for churn forecasting and component identification in the telecom sector,” *J. Comput. Biomed. Inform.*, vol. 7, no. 1, pp. 264–280, 2024.
[46] H. Farman, D. Khan, S. Hassan, M. Hussain, and S. A. A. Usmani, “Analyzing machine learning models for forecasting precipitation in Australia,” *J. Comput. Biomed. Inform.*, vol. 7, no. 1, pp. 439–458, 2024.
[47] R. Happle, K. J. Kalveram, U. Büchner, K. Echternacht-Happle, W. Göggelmann, and K. H. Summer, “Contact allergy as a therapeutic tool for alopecia areata: application of squaric acid dibutylester,” *Dermatology*, vol. 161, no. 5, pp. 289–297, 1980.
[48] H. Farman, S. Ahmed, M. Imran, Z. Noureen, and M. Ahmed, “Deep learning based bird species identification and classification using images,” *J. Comput. Biomed. Inform.*, vol. 6, no. 1, pp. 79–96, 2023.
Downloads
Published
How to Cite
Issue
Section
License
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (CC-By) that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
This work is licensed under a Creative Commons Attribution License CC BY
