Authors :
Dr. R Amutha; A.S. Rohith Raghav; Amulya C.; Jyothi U.; Keerthana M.G.
Volume/Issue :
Volume 10 - 2025, Issue 1 - January
Google Scholar :
https://tinyurl.com/43fd78um
Scribd :
https://tinyurl.com/y2x5s3ru
DOI :
https://doi.org/10.5281/zenodo.14792171
Abstract :
The system utilizes state-of-the-art machine learning algorithms and deep learning techniques to analyze facial
features and predict BMI with high accuracy. The primary advantage of using facial recognition technology is its non-
invasive nature, which ensures a user-friendly and accessible approach to health monitoring. This method can make benefits
for individuals with physical disabilities, as it eliminates the cumbersome or intrusive measurements. This model to trained
on dataset that includes individuals with a single or more physical disabilities. The dataset encompasses various facial
features and annotations that are pertinent to BMI prediction. By incorporating this diverse data, the model is made robust
and capable of handling the unique variations in the physically challenged population.
The process initiates with image capturing, where users can either upload an existing image or capture a new one using
a camera. The system then performs detection of facial landmarks to identify facial features, such as the eyes, nose, mouth,
and facial contour. These landmarks are crucial extraction of meaningful features that correlate with BMI. The detected
facial landmarks are then normalized to ensure consistency across different images, mitigating variations due to scale,
orientation, and lighting conditions. Subsequently, the normalized facial points are used to extract relevant features. Few
features may include distances between specific landmarks, ratios, and geometric shapes formed by the landmarks.
Advanced feature extraction techniques and statistical analyses are employed to derive these metrics. The extracted features
serve as input to the predictive model, which has used supervised learning for training. Once the training is done, it
is capable of predicting BMI from facial features with high precision. The system provides real-time feedback, allowing
users to monitor their BMI effortlessly. The non-invasive nature of this approach makes it ideal for routine health
monitoring and early detection of potential health risks.
Keywords :
Body Mass Index (BMI), Facial recognition, Disabilities, Machine Learning (ML).
References :
- Yujin Wang; Zhi Jin; Jia Huang; Hongzhou Lu; Wenjin Wang, “Facial Landmark based BMI Analysis for Pervasive Health Informatics”, 2023.
- B. SrinivasaRao; Y. Ashok Kumar; V. Vinay Sai; G. Srinu, “BMI Estimation via Facial ImageAnalysis” 2024.
- ZhiJin; Junjia Huang; Wenjin Wang; Aolin Xiong; Xiaojun Tan, “Estimating Human Weight From a Single Image”, 2024.
- Chong Yen Fook; Lim Chee Chin; Vikneswaran Vijean, “Investigation on BMI Prediction from Face Images”, 2020.
- Facial Landmark based on BMI Analysis for Pervasive Health Informatics, 2023.
- BMI Estimation via Facial Image Analysis, 2024.
- Estimating Human Weight from a Single Image, 2024.
- Investigation on Body Mass Index (BMI) Prediction from Face Images, 2020.
The system utilizes state-of-the-art machine learning algorithms and deep learning techniques to analyze facial
features and predict BMI with high accuracy. The primary advantage of using facial recognition technology is its non-
invasive nature, which ensures a user-friendly and accessible approach to health monitoring. This method can make benefits
for individuals with physical disabilities, as it eliminates the cumbersome or intrusive measurements. This model to trained
on dataset that includes individuals with a single or more physical disabilities. The dataset encompasses various facial
features and annotations that are pertinent to BMI prediction. By incorporating this diverse data, the model is made robust
and capable of handling the unique variations in the physically challenged population.
The process initiates with image capturing, where users can either upload an existing image or capture a new one using
a camera. The system then performs detection of facial landmarks to identify facial features, such as the eyes, nose, mouth,
and facial contour. These landmarks are crucial extraction of meaningful features that correlate with BMI. The detected
facial landmarks are then normalized to ensure consistency across different images, mitigating variations due to scale,
orientation, and lighting conditions. Subsequently, the normalized facial points are used to extract relevant features. Few
features may include distances between specific landmarks, ratios, and geometric shapes formed by the landmarks.
Advanced feature extraction techniques and statistical analyses are employed to derive these metrics. The extracted features
serve as input to the predictive model, which has used supervised learning for training. Once the training is done, it
is capable of predicting BMI from facial features with high precision. The system provides real-time feedback, allowing
users to monitor their BMI effortlessly. The non-invasive nature of this approach makes it ideal for routine health
monitoring and early detection of potential health risks.
Keywords :
Body Mass Index (BMI), Facial recognition, Disabilities, Machine Learning (ML).
