Project detail

Latte Lies And Milky Myths: Dectecting Adulterants With Machine Learning

Project goal

This was my science fair project for 10th grade. I coded a machine learning model to detect adulterants in milk and coffee grains using just an image.

Why it matters

This project matters because adulteration is when companies add cheap or dangerous substances to food products, and it can lead to serious health issues. Such as gastrointestinal problems, cancer, allergic reactions, and more. It is most prominent in India although it occurs worldwide. One example in the US was when perchlorate was found in milk in 2004 in states like California, Texas and Georgia among others. There have also been cases of milk adulteration with chemicals like formaldehyde and urea. These are extreme cases but the most common adulterants for milk specifically are water, sugar, salt and starch. For more info I will attach a pdf of my trifold board below.

What I improved

If I continue this project, then I will use more images and more diverse images for the training dataset in order to increase the accuracy of the model.

Awards

  1. Top 10 at the Harrison High School Science And Engineering Fair (Qualifies for Regional Competition)
  2. Earned a Office of Naval Research - Naval Science Award, at the Harrison High School Science And Engineering Fair
  3. Earned 2nd Place at the Cobb/Paulding Regional Science Fair in my category