VITAMIN DEFICIENCY DETECTION USING DEEP LEARNING

Abstract

This study presents a deep learning-based system for automated detection of vitamin deficiencies using medical image analysis. Vitamin deficiencies are widespread nutritional disorders that can lead to severe health complications if left undiagnosed, making timely and accurate detection crucial for preventive healthcare. Traditional diagnostic methods rely on invasive biochemical tests, which are time-consuming, costly, and often inaccessible in low-resource settings. To address these challenges, we developed a convolutional neural network (CNN) model capable of classifying vitamin deficiency types directly from medical images, such as photographs of skin, nails, eyes, or other visible clinical signs. The model was trained and evaluated on a dataset curated from publicly available medical image repositories and clinical sources, incorporating data augmentation techniques to improve generalization. Experimental results demonstrate that the proposed model achieves a classification accuracy of approximately 95%, with high precision, recall, and F1-scores across multiple deficiency classes. The system exhibits strong potential as a non-invasive, rapid, and cost-effective tool for preliminary vitamin deficiency screening, particularly in resource-constrained environments. Future work aims to expand the dataset size, integrate multimodal data, and deploy the solution in mobile or telemedicine applications to enhance accessibility and real-time diagnostics.