How accurate are photo body-fat estimates?

Short answer: accurate enough to show a trend, not accurate enough to argue with a clinic. That is a real distinction and most apps blur it.

What a photo scan is doing

It measures silhouettes, not tissue. From a front and a side photo, with your height as the scale reference, it estimates circumferences — waist, hip, chest, thigh and so on — and derives a body-fat figure from those proportions.

Nothing in that pipeline sees fat. It sees shape, and infers.

The realistic error

In good conditions — even lighting, fitted clothing, a plain background, a still pose — circumference estimates land within roughly one to three centimetres.

Derived body-fat percentages carry more error than the circumferences they come from, because a small error in the waist becomes a larger error once it is run through a formula.

Why a trend is still worth having

A consistent method with a known bias is useful even when it is not exact. If the estimate is two centimetres high every month, the change between months is still real.

This is the same reason a bathroom scale is useful despite water weight: you are reading the direction, not the absolute.

How to make yours less wrong

Same time of day, same lighting, same clothing, same distance from the camera.

Fitted clothing, not loose. A t-shirt hides the waist the model is trying to find.

Stand still and straight, arms slightly away from the body.

Correct anything you know is wrong before you save it — a value you measured with a tape beats a value inferred from a photo.

When to use something else

If a number is going to inform a medical decision, get a clinical measurement. DEXA is the practical reference standard; hydrostatic weighing and air-displacement plethysmography are alternatives. A phone is not in that category and no honest app will claim it is.

Questions