Issuu

International Research Journal of Engineering and Technology (IRJET)

e-ISSN: 2395-0056

Volume: 10 Issue: 06 | Jun 2023

p-ISSN: 2395-0072

www.irjet.net

PREDICTION OF BMI FROM FACIAL IMAGE USING DEEP LEARNING V.Divya Raj1, Shreya Motkar2, N.Yamini Swamy3, N. Sanjana4, Shereen Sultana5 1Asst. Professor, Computer Science and Engineering, GNITS, Hyderabad, Telangana, India

2345 UG Students, Computer Science and Engineering, GNITS, Hyderabad, Telangana, India

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - Any person's BMI (Body Mass Index) is an

cardiovascular and other hazardous diseases increases with increasing BMI. On the other hand, some people struggle with issues like inadequacies and malnutrition. There are primarily four classifications based on BMI values: underweight (BMI <18.5),normal(18.5<BMI≤25),overweight(25<BMI<30), and obese(30<BMI), BMI can therefore assist a person in keeping track of their health.

important sign of their health. It determines whether the person is underweight, normal weight, overweight, or obese. It is a gauge of a person's health in relation to their body weight. Typically, the centre and lower parts of the face are wider on fat people. Without a measuring tape and a scale, it is challenging for the person to calculate their BMI. This method uses deep learning and transfer learning models like VGG-Face, Inception-v3, VGG19, and Xception to discover a correlation between BMI and human faces in order to develop a strategy that predicts BMI from human faces. Three publicly accessible datasets (Arrest Records Database, VIP-Attribute Dataset, and Illinois DOC dataset) including pictures of both inmates and Hollywood celebrities were used to create the front-facing photos. The pictures could be blurry, irregular, or even have titles. A technique known as StyleGan is used to make them look identical. The face is then vertically aligned using a face landmark detection model called DLIB 68, and the background is blurred to isolate the face. The pre-trained model’s whole network of completely linked layers is added, and the result is the person's BMI. The existing methodology differs significantly from the current system in that it makes use of pre-trained models like Inception-v3, VGG-Face, which employs computer vision to improve performance and shorten training time.

A person can be learned numerous things just by looking at their face. Recent research has demonstrated a significant relationship between a person's BMI and their facial features. People with thin faces are likely to have lower BMIs, and vice versa. Obese people typically have bigger middle and lower facial features. If the person does not have a measuring device and a scale, calculating BMI can be challenging. Deep learning has made tremendous strides recently, enabling models to extract useful information from photos. These techniques allow us to extrapolate the BMI from human faces. Therefore, we have suggested a method to predict BMI from human faces in this study. This technique might make it easier for health insurance firms to keep track of their clients' medical histories. Additionally, the government might monitor the health statistics of a certain area and create laws in accordance with them.

2. OBJECTIVES

Key Words: Body Mass Index prediction, Face To BMI, Deep Learning, Facial Features, StyleGan, DLIB 68.

The proposed system's goals are as follows:

1.INTRODUCTION

• To improve image quality by pre-processing the inconsistent images in the dataset.

Body Mass Index (BMI) is a commonly used index that uses the ratio of a person's height to weight to reflect their general weight condition. Numerous aspects, including physical health, mental health, and popularity, have been linked to BMI. BMI calculations frequently call for precise measurements of height and weight, which entail laborintensive manual labour. Any person's BMI (Body Mass Index) is an important sign of their health. If the person is underweight, normal, overweight, or obese, it is determined. Health continues to be one of the most overlooked factors. Even technology with many advantages has its downsides. It has made people more slothful, which has decreased their physical activity and resulted in a sedentary lifestyle and an increase in BMI, both of which are harmful to their health and raise the risk of chronic diseases. The likelihood of acquiring

Impact Factor value: 8.226

• To align the face vertically when the image is tilted. • To predict BMI by extracting facial traits from images. • To use deep learning to predict Body Mass Index from previously pre-processed facial images.

3. LITERATURE SURVEY 1.‘’A computational approach to body mass index prediction from face images” by Lingyun Wen and Guodong Guo (2013): In this study, the BMI was determined computationally. By employing the Active Shape Model to extract facial landmarks from facial images, the authors were able to extract seven facial traits. These seven characteristics include ES (Eye Size), CJWR

ISO 9001:2008 Certified Journal

Page 671