Skin Allergy Detection Method using Computer Aided Design

The biomedical industry is facing a major challenge in skin allergy detection and monitoring. The increasing contamination levels and consumption of unhealthy food have led to a higher number of individuals experiencing skin-related problems. As a result, the incidence of patients with skin-related issues is expanding rapidly. Nevertheless, traditional diagnostic approaches heavily rely on subjective assessments by healthcare experts, leading to inconsistencies and delays in the diagnosis process. In this work, we employed image processing and deep learning (DL) techniques to detect skin allergies in various images and evaluate the accuracy and efficiency of the methods used. The Support Vector Machine (SVM), as well as DL models including AlexNet, ResNet50, and VGG16, accomplished the detection of skin allergies in the images. A total of 2609 images, encompassing 19 different classes of skin allergy images, were utilized in this work. These images contained various types of skin allergies, such as acne, atopic dermatitis, bacterial infections, cancerous lesions, drug eruptions, eczema, lichen, melanoma, skin keratosis, and other classes. The results proved that the accuracy of the ResNet50 method was 98%, and the sensitivity value was 100% in detecting skin allergies in RGB images. From the findings, it can be said that ResNet50 outperformed the other methods employed in this work for skin allergy detection. Moreover, the accuracy and sensitivity values obtained in this work surpassed those reported in previous research studies within the same field.