An Analysis of Alopecia Areata Classification Framework for Human Hair Loss Based on VGG-SVM Approach

Abstract

Artificial intelligence approach is used in this article to make a diagnosis of hair loss. An autoimmune condition known as alopecia areata (AA) results in hair loss in the affected area. The most recent figures show that AA has an incidence of 2% and a frequency of 1 in 1000 people worldwide. For instance, classification is important in the field of medicine because one of the doctor's main objectives is to establish whether or not a patient has a condition. The identification of alopecia areata may be helpful for better prediction and diagnosis. Machine learning (ML) techniques have shown potential in a variety of domains, including dermatology. The dynamic character of illness symptoms is another important factor in the precise diagnosis of a particular disease. After then, as was already mentioned, linked work in the fields of ML and classification of healthy hair & hair disorders. The purpose of this research is to evaluate the efficacy of neural networks in recognising alopecia and non-humans. Healthy hair (HHs) and alopecia areata (AA) have been classified using a framework that will be susceptible to IP, including CLAHE enhancement and segmentation. Then, to increase the precision of the proposed framework, data augmentation (DA) was employed to generate further data. The VGG- 19 pre-trained CNN model was then used to extract features. The Support Vector Machine (SVM) classification approach is then used to create an ML model utilising 70% of the images. The remaining images in the collection were used for testing. The suggested VGG-SVM was demonstrated to be 98.31% accurate in the simulation, which used 200 images of HH from the Figaro1k dataset and 68 images of AA from the Dermnet dataset. The proposed VGG-SVM outperformed the current Edge-SVM in terms of accuracy. The findings of our investigation point to the potential for enhanced dermatological prediction.