Detecting Body Mass Index from a Facial Photograph in Lifestyle Intervention

Abstract

This study aimed to identify whether a research participant’s body-mass index (BMI) can be correctly identified from their facial image (photograph) in order to improve data capturing in dissemination and implementation research. Facial BMI (fBMI) was measured using an algorithm formulated to identify points on each enrolled participant’s face from a photograph. Once facial landmarks were detected, distances and ratios between them were computed to characterize facial fatness. A regression function was then used to represent the relationship between facial measures and BMI values to then calculate fBMI from each photo image. Simultaneously, BMI was physically measured (mBMI) by trained researchers, calculated as weight in kilograms divided by height in meters squared (adult BMI). Correlation analysis of fBMI to mBMI (n = 1210) showed significant correlation between fBMI and BMIs in normal and overweight categories (p < 0.0001). Further analysis indicated fBMI to be less accurate in underweight and obese participants. Matched pair data for each individual indicated that fBMI identified participant BMI an average of 0.4212 less than mBMI (p < 0.0007). Contingency table analysis found 109 participants in the ‘obese’ category of mBMI were positioned into a lower category for fBMI. Facial imagery is a viable measure for dissemination of human research; however, further testing to sensitize fBMI measures for underweight and obese individuals are necessary.

Facial Adiposity, Attractiveness, and Health: A Review

“…One of the key features of the Wize Mirror is the use of facial morphometric analysis to predict cardio-metabolic risk. Studies by Kocabey et al (2017) and Barr et al (2018) also reported on the development of Face-to-BMI- Systems that predicts BMI from facial images found on social media platforms, for example. The results reported by these studies indicate that these computer models could predict actual BMI from facial cues alone to a degree of accuracy very similar to that of human observers.…”

“…Another study by Re and Rule (2016b) corroborated this finding by reporting that an average change in BMI of 1.33 kg/m 2 was sufficient for participants to report a noticeable change in the appearance of faces. In recent years there has also been a concerted effort from researchers to develop computer vision methods (Wen and Guo, 2013; Kocabey et al, 2017; Barr et al, 2018) and statistical models (Wolffhechel et al, 2015; Stephen et al, 2017) to predict BMI from facial images.…”

Body Mass Index Prediction and Classification Based on Facial Morphological Cues Using Multinomial Logistic Regression

“…This paper is not concentrating more on establishing the relationship between the facial fiducial points and BMI. Because the previous studies [1,8,9,12,21] proved the existence of the relationship between the facial cues and the BMI; instead, this paper straight away goes to the prediction of BMI [22], by which, it convinces the objective of the study that the MLR model yields better results than the other methods. Hence, the facial features, such as geometrical and texture, extracted from each facial image [23,24] were straight away subject to the experiments using the MLR model discussed in Section 3.…”

“…They have applied multivariate linear regression to estimate the association between facial shape and texture with BMIand WHR. Barr et al [12] have attempted to examine whether the existing methods correctly identifies the facial cues and BMI. To arrive the above objective, they have deployed the regression to assess the relationship between the facial cues and BMI; further, they have used the Correlation and Contingency table analyses and reported that facial image features are a viable measure for the dissemination of human research related to facial cues and BMI, although, they suggested that it requires further analysis.…”

Body Weight Estimation using 2D Body Image

“…An Active Shape Model is used to extract facial features which are used to predict BMI using various regression techniques. Barr et al [8] used facial landmarking to figure out adiposity (facial fatness) which positively correlates to the weight of the person. This method is less acurate in exteme underweight and obese cases though.…”