Convolutional and Graph Neural Networks for the prediction of Papulosquamous Diseases (PSDs)

The precise identification of papulosquamous diseases (PSDs), including psoriasis and seborrheic dermatitis, is a challenge for dermatologists due to their complicated and heterogeneous clinical presentations. In the discipline of dermatology, artificial intelligence (AI) has recently made major advancements in image classification and malignancy prediction. Deep learning is becoming increasingly popular for treating a variety of diseases. This study proposes a CNN (Convolutional Neural Networks)+GNN (Graph Neural Networks) architecture and evaluates and compares the performance of multiple deep learning models, including LSTM (Long Short-Term Memory), Autoencoders (AEs), and Recurrent Neural Networks (RNNs), for the automatic categorization of Papulosquamous Diseases(PSDs). . A hybrid strategy combining CNN and GNN was selected as the most efficient model, achieving the maximum performance across all measures with 93% accuracy, 92% precision, 94% recall, and 93% F1-Score.