Systemic Lupus Erythematosus (SLE) Detection Using Deep Learning

Systemic lupus erythematosus (SLE) is a chronic autoimmune disease with diverse clinical manifestations, making early diagnosis challenging. This study investigates the efficacy of convolutional neural networks (CNNs) for detecting Systemic Lupus Erythematosus (SLE) using deep learning methodologies. Three CNN architectures, namely Dilated-CNN, CNN with Attention Mechanism (CNN-AM), and standard CNN, were evaluated on their ability to distinguish between SLE-positive and SLE-negative cases based on dermatological images from the DermNet dataset. The models were trained and compared using various performance metrics, including accuracy, precision, recall, F1-score, and ROC AUC. The results revealed that CNN-AM achieved the highest accuracy (0.96651) and outperformed Dilated-CNN (0.8795) and standard CNN (0.9257) in all evaluation metrics. This study highlights the potential of CNNs, particularly those incorporating attention mechanisms, for accurate and reliable SLE detection using deep learning, paving the way for further research and development of improved diagnostic tools for SLE patients.