Recent Papers

Abstract-Traditional Body Mass Index (BMI) estimation requires height and weight measurements, which are often impractical in remote settings. This paper presents a complete ML-based BMI prediction system using facial recognition, integrating face detection, preprocessing, feature extraction, and multiple ML models including Random Forest, SVM, CNN, ResNet, MobileNet, and Vision Transformer (DINOv2). The proposed system achieves high accuracy and supports telemedicine, fitness applications, and public health screening.Keywords: - BMI Estimation, Machine Learning, Facial Recognition, CNN, Res Net, Mobile Net, Vision Transformer.IntroductionBody Mass Index (BMI) is a widely used metric to assess a person's body fat based on their height and weight. It serves as a crucial indicator for identifying potential health risks such as obesity, cardiovascular diseases, and diabetes. Traditional methods of BMI calculation require direct physical measurements, which may not always be feasible in remote or large-scale settings. Recent advancements in Artificial Intelligence (AI) and Machine Learning (ML), This project explores an innovative approach to estimating an individual's BMI using facial recognition technology. This project aims to leverage Machine Learning (ML) and Facial Recognition to estimate BMI from facial features alone. Research suggests that specific facial characteristics such as face shape, cheekbone prominence, jawline structure, and fat distribution can correlate with BMI. Using these insights, the project proposes a system that can analyze facial images and accurately predict BMI levels, eliminating the need for direct physical data. Such a solution has significant applications in remote healthcare, fitness monitoring, digital health apps, and public health screenings where physical access to patients or users is limited. The use of ML algorithms ranging from traditional methods like Support Vector Machines (SVM) and Random Forest to deep learning models like Convolutional Neural Networks (CNN), Res Net, and Mobile Net is explored and compared for optimal accuracy and performance. This innovative, non-invasive approach not only aligns with modern healthcare trends but also promotes accessibility, privacy, and efficiency in health monitoring systems. In today's data-driven world, advancements in artificial intelligence and computer vision have opened new possibilities form on contact, fast and assessments. The project focuses on ML-based BMI prediction using facial recognition, which applies machine learning and computer vision techniques to estimate an individuals Body Mass Index (BMI) from facial images. By using Convolutional Neural Networks (CNNs) for feature extraction including facial shape, texture, and contour the system provides a non-invasive and automated method for BMI estimation without requiring height or weight data. The predictive model is trained on a dataset of facial images with known BMI values, enabling accurate BMI prediction for new individuals. This AI-driven healthcare approach supports health monitoring, fitness assessment, and early detection of obesity or undernutrition, promoting personalized health recommendations and advancing modern medical applications.

Copyright © 2025 Manoj Dyavannavar, PruthviRaj C, RaviKumar N J, Sharanabasava. This is an open access article distributed under the Creative Commons Attribution License.