Problem

The problem we aimed to solve was the high rates of malnutrition in developing countries, particularly among children under 5 years of age. Malnutrition can lead to stunted growth and development, as well as other health issues. Traditional methods of measuring a child's nutritional status involve physical measurements and can be time-consuming and costly. Our goal was to create a more efficient and accessible solution using artificial intelligence.

We identified that there was a need for an accurate and easy-to-use tool that could predict a child's height, weight, and nutritional status using only a picture. This would allow for quicker identification of at-risk children and enable healthcare professionals to intervene earlier.

Learn more about our project here.

Solution

Our team developed an artificial intelligence model that could accurately predict a child's height, weight, and nutritional status based on a picture of them. We used machine learning algorithms and computer vision techniques to analyze the images and extract relevant features such as facial landmarks, and body proportions.

We trained our model on a large dataset of images from children in developing countries with known height, weight, and nutritional status. We also incorporated demographic information such as age, gender, and location into our model to improve its accuracy.

Impact

Reduced the time it takes to assess a child's nutritional status.
Percentage Relative error predicting height is 3.4%
Percentage Relative error predicting weight is 9.1%

"Now, using unconstrained images of a child, we can already have a first glimpse on his the nutritional status ! "

Team

Team Member #1 - Ahmed Ben Haj Yahia - Machine Learning Engineer
- Developed and trained the artificial intelligence model
- Analyzed and processed the image data using computer vision techniques
Team Member #2 - Thomas Grégoire Stegmüller - PhD and my supervisor

AI against malnutrition