NutriVision: Early Detection of Malnutrition Using Advanced Deep Learning

Nowadays, malnutrition has become a major public health issue with profound consequences for human health and well-being. In the case of the treatment and prevention of malnutrition, accurate and timely risk detection is crucial to adopt the appropriate action or receive the required treatment. Identifying those who are malnourished can help lower the risk of mortality as well as physical and health problems. NutriVision is a deep learning-based mobile application for early detection of malnutrition in children using image-based analysis. Based on an image dataset collected from the People's Primary Healthcare Initiative (PPHI), Paramedical Staff of child health images and medically relevant contextual data. The app uses an image-driven Convolutional Neural Network (CNN) model. The model achieved an accuracy of 97% on the test set and had a precision score, recall rate and F1 score. NutriVision is a non-invasive, cost-effective, and scalable solution for point-of-care malnutrition detection that offers the potential as an effective technology to combat the health problems seen in children, especially within low-resource settings.