Body composition estimates using a 2D image analysis system across different environmental conditions: An agreement study
Related Experiment Video
Updated: Nov 8, 2026
Segmentation and Linear Measurement for Body Composition Analysis using Slice-O-Matic and Horos
Published on: March 21, 2021
Casey J Metoyer1, Katherine Sullivan2, Lee J Winchester3
1Notre Dame Sports Performance, The University of Notre Dame, Notre Dame, Indiana, USA.
Purpose:
This study examined the agreement between %Fat measurements using a smartphone-based application (IMAGE) across different environmental conditions.
Methods:
A single reference image was obtained using an 8 MP smartphone camera under Ambient Light in front of a white background. Additional photos were obtained using a 0.7 MP, 5 MP, and 12 MP smartphone cameras; low-, moderate-, and bright-lighting conditions; and various color backgrounds including black, green, orange, and gray.
Results:
%Fat measured using the 0.7 MP camera (27.8 ± 6.2 %Fat) was higher than the reference (26.8 ± 6.1 %Fat) (p < 0.001). The black (32.0 ± 12.0 %Fat), green (27.5 ± 6.3 %Fat), and gray (27.8 ± 6.3 %Fat) backgrounds yielded higher %Fat than the white (p = 0.03, 0.01, and 0.001). All camera, lighting, and background conditions were strongly correlated with the reference (all intraclass correlation coefficient [ICC] >0.98, all standard error of the estimate [SEE] <1.5 %Fat, all p < 0.001), except the black background which yielded poorer agreement with the white background (ICC = 0.69, SEE = 4.5%, p < 0.001).
Conclusion:
%Fat from IMAGE were strongly correlated across various environmental conditions.
