English VersionVersione italiana  – Versión en español

The body composition estimates are computed from the picture (digital image) of a lateral standing upright profileconditioned to a black and white profile. Therefore, the image must be as free as possible of position and environmental influences to obtain an optimal and correct count of the black and the white pixels representative of the individual. To enable these objectives, I provide some “must do” recommendations.

  1. The picture should be well illuminated avoiding lateral lamps projecting significant shades
  2. The lateral profile of the subject must show only the body’s profile: arms kept stretched (extended) vertically without protruding anterior and posterior silhouette boundaries. Legs fully extended, and feet well visible
  3. The subject stands upright or fully erect at attention (similar to the soldier shown below) with eyes focused straight ahead.

These pictures show adequate postures. Note that the bulk of the outfit will become the detected surface of silhouette, altering the real fat mass estimates. the real silhouette. Elasticated bodysuit or swimsuits are ideal.

Minimal dressing or tight garments not altering the real profile is ideal.

Exploiting Deep and Machine Learning and Neural Network allows pictures taken with any background. Human body profiles are automatically detected and conditioned immediately to black & white silhouette

The smart phone must be held halfway of the subject’s height, and vertical.

a silhouette of a woman holding a stick

The use of our “FitYourOutfit” APP helps obtaining correct aiming, distance and parallaxes.

To ensure valid body composition estimates, enter actual personal data (stature or standing height and weight) at the time the picture (image) is obtained. Best results achieved if stature values are entered in metric system, as feet and inches have larger approximations (1 cm = ¼ in). Breathe normally and avoid contractingabdominal muscles whilst taking a picture.

MOST COMMON MISTAKES: DON’Ts !