Pictures to support a screening decision.

What the phone is intended to do

The capture sequence asks for front, side and back views of one child. The planned height estimate can be considered with age and sex for height-for-age screening. For wasting, GrowthSight is exploring two separate routes: use its height estimate with weight from a Bluetooth-connected scale, or estimate mid-upper arm circumference (MUAC) from the same images.

The goal is to give a worker a measurement to review while the child and caregiver are still present. A result should inform the next step under the local screening pathway; it should not be treated as a diagnosis on its own.

What AI cannot establish by itself

An image estimate cannot replace a full nutritional assessment, clinical judgement or checks required by a programme. Image conditions, the child’s pose and reference measurements all matter when evaluating whether a tool is reliable in the field.

GrowthSight has not published field validation or clinical performance figures on this website. The product remains in development, and pilot work is intended to test measurement quality, workflow fit and how trained workers use the findings.

What to ask before a pilot

A useful pilot question is specific: which children will be screened, which reference measurements will be collected, how will poor-quality captures be handled, and who will decide on referral or follow-up? These questions matter more than an unsupported claim that AI alone can detect malnutrition.