Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multimodal AI for Body Fat Estimation: Computer Vision and Anthropometry with DEXA Benchmarks

Rayan Aldajani

TL;DR

The paper tackles the challenge of affordable body fat estimation by AI, addressing the inaccessibility of DEXA and the inaccuracies of simple anthropometric methods. It introduces two approaches—a CNN-based image regression and an anthropometric regression—evaluated on two separate datasets (282 frontal images with self-reported BF% and 253 anthropometric records). The image-based CNN achieves RMSE $= 4.44\%$ and $R^2 = 0.81$, outperforming the anthropometric baseline, indicating AI can provide low-cost BF% estimates. The work lays the groundwork for multimodal fusion by sharing a dataset and demonstrating feasibility, while outlining future directions to improve generalization with larger, diverse, paired datasets and privacy-preserving deployment for at-home use.

Abstract

Tracking body fat percentage is essential for effective weight management, yet gold-standard methods such as DEXA scans remain expensive and inaccessible for most people. This study evaluates the feasibility of artificial intelligence (AI) models as low-cost alternatives using frontal body images and basic anthropometric data. The dataset consists of 535 samples: 253 cases with recorded anthropometric measurements (weight, height, neck, ankle, and wrist) and 282 images obtained via web scraping from Reddit posts with self-reported body fat percentages, including some reported as DEXA-derived by the original posters. Because no public datasets exist for computer-vision-based body fat estimation, this dataset was compiled specifically for this study. Two approaches were developed: (1) ResNet-based image models and (2) regression models using anthropometric measurements. A multimodal fusion framework is also outlined for future expansion once paired datasets become available. The image-based model achieved a Root Mean Square Error (RMSE) of 4.44% and a Coefficient of Determination (R^2) of 0.807. These findings demonstrate that AI-assisted models can offer accessible and low-cost body fat estimates, supporting future consumer applications in health and fitness.

Multimodal AI for Body Fat Estimation: Computer Vision and Anthropometry with DEXA Benchmarks

TL;DR

The paper tackles the challenge of affordable body fat estimation by AI, addressing the inaccessibility of DEXA and the inaccuracies of simple anthropometric methods. It introduces two approaches—a CNN-based image regression and an anthropometric regression—evaluated on two separate datasets (282 frontal images with self-reported BF% and 253 anthropometric records). The image-based CNN achieves RMSE and , outperforming the anthropometric baseline, indicating AI can provide low-cost BF% estimates. The work lays the groundwork for multimodal fusion by sharing a dataset and demonstrating feasibility, while outlining future directions to improve generalization with larger, diverse, paired datasets and privacy-preserving deployment for at-home use.

Abstract

Tracking body fat percentage is essential for effective weight management, yet gold-standard methods such as DEXA scans remain expensive and inaccessible for most people. This study evaluates the feasibility of artificial intelligence (AI) models as low-cost alternatives using frontal body images and basic anthropometric data. The dataset consists of 535 samples: 253 cases with recorded anthropometric measurements (weight, height, neck, ankle, and wrist) and 282 images obtained via web scraping from Reddit posts with self-reported body fat percentages, including some reported as DEXA-derived by the original posters. Because no public datasets exist for computer-vision-based body fat estimation, this dataset was compiled specifically for this study. Two approaches were developed: (1) ResNet-based image models and (2) regression models using anthropometric measurements. A multimodal fusion framework is also outlined for future expansion once paired datasets become available. The image-based model achieved a Root Mean Square Error (RMSE) of 4.44% and a Coefficient of Determination (R^2) of 0.807. These findings demonstrate that AI-assisted models can offer accessible and low-cost body fat estimates, supporting future consumer applications in health and fitness.

Paper Structure

This paper contains 14 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. The Importance of Accessible Body Fat Estimation
  3. Limitations of Current Methods
  4. Role of Artificial Intelligence
  5. Contribution of This Study
  6. Background
  7. Anthropometric Models
  8. Gold-standard Imaging Techniques
  9. Artificial Intelligence in Health Applications
  10. Results
  11. Dataset Description
  12. Anthropometric Model Results
  13. Image-Based Model Results
  14. Discussion and Conclusion

Figures (2)

  • Figure 1: Scatter plot of predicted vs. true body fat percentage for the anthropometric regression model (using weight, chest, abdomen, hip, and thigh). The $R^2$ value of 0.57 indicates moderate predictive ability.
  • Figure 2: Training and validation loss curves for the image-based CNN model over 50 epochs. The model converges stably, with training loss decreasing steadily and validation loss plateauing at a higher level.