Inspiration
The ToxiScore is inspired by two SDGs for long term health and wellbeing.
What it does
The ToxiScore gives you an indicator of how toxic the ingredients of a food product is for your health. Usually it is very hard to understand and identify the impact of food ingredients. A Nutriscore is giving you an indicator of how nutral and good the ingredients are but it does not tell you about dangerous ingredients which might impact individual people's health negative. It categorizes the product in green - low danger, yellow - medium danger and red - high danger.
How we built it
The idea is to have an app which can extract the text from a package through a picture. The data is mapped to predefined categories by maschine learning. The list of ingredients for the mapping is extending through maschine learning process. If ingredients fall into category danger the red score will be given out. The final score can also be adjusted by an individual health profile to be more precise.
Challenges we ran into
Challenges were first the overall extraction of text from a photo and comparability of several systems. The code of matching ingredients to categories and giving a score is also still in development and creating programming issues.
Accomplishments that we're proud of
We designed categories and a logic. We are happy with the brand, label and benefits that it could bring.
What we learned
We learned how complex is the usability of several technical steps to create an automated process.
What's next for ToxiScore
The code needs to be completed and finalized.
Built With
- nannowe
- ocr
- phython
Log in or sign up for Devpost to join the conversation.