Body composition assessment using two-dimensional digital image analysis

A method of deriving body composition of an individual, the method comprising:

• obtaining, via a body weight scale, a body weight measurement of the individual and an impedance of the individual;
• obtaining, via a camera sensor of a mobile device, a single image of the individual comprising an anterior or posterior portion of a body of the individual with hips and shoulders parallel to the camera sensor;
• digitizing, by a processor of the mobile device, the single image of the individual into an array of pixels;
• identifying, by the processor and using a digital anthropometer on the mobile device, a plurality of anatomical landmarks in the digitized single image, including at a waist landmark, a hip landmark, and a shoulder landmark;
• measuring, by the processor and using an automated image analysis program that is configured to measure the array of pixels, a cross-sectional diameter as a linear distance of pixels in the single image at each of the plurality of anatomical landmarks;
• estimating, by the processor, a body composition value of the individual by applying a multivariate linear regression model configured for the plurality of anatomical landmarks;
• determining, by the processor, a body weight of the individual from the body weight measurement from the body weight;
• determining, by the processor, total body water of the individual using bioelectrical impedance analysis (BIA) of the impedance obtained from the body weight scale;
• estimating, by the processor, body fat of the individual using an estimated body volume, the body weight, and the total body water;
• and outputting, by the processor, the body composition value to a digital display on the mobile device via an application executing on the mobile device,
• wherein the body composition value displayed on the mobile device is used to monitor changes to the individual's body composition over time;
• wherein the mobile device is one or more of a smartphone and a tablet;
• wherein the single image is a single 2D image and where the identifying of the marks occurs only on the single 2D image.