Diagnosis of psoriasis and lichen planus in real-time using neural networks based on skin Biomechanical properties obtained from numerical simulation.

Due to their similar clinical presentations, the scarcity of competent dermatologists, and the urgency of diagnosis, the accurate diagnosis of dermatological conditions such as Psoriasis and Lichen Planus is challenging. This study introduces a novel approach leveraging deep learning and numerical simulations of skin biomechanical properties to enhance diagnostic precision. By utilizing ABAQUS software, this study incorporates 1000 numerical simulations for data generation to combat the limitations of datasets in these ailments. Utilizing the ResNet-50 convolutional neural network (CNN), this research integrates data from finite element simulations based on variations in skin's biophysical parameters. The dataset comprises 1000 instances, evenly divided between Psoriasis and Lichen Planus, with attributes including displacement, humidity, age, and sex. The numerical data were converted into image data to optimize the ResNet-50 model's performance. The results were validated through 5-fold cross-validation, 3-fold cross-validation, and random splitting. The proposed methodology demonstrated remarkable diagnostic accuracy, achieving 99.8% using 5-fold cross-validation, surpassing previous investigations, and highlighting the potential of combining AI and biomechanical simulations for real-time skin disease diagnosis to assist physicians and dermatologists in classifying skin diseases.