Inspiration
We know what it's like to feel like you don't have options when it comes to cooking.
What it does
FoodAllert is a cloud based web application that takes images of recipes from recipe books or websites and stores them in the Google Cloud. It then uses Google Vision API to perform text detection on the image. Natural language processing is used to extract the ingredients from the recipe and to classify them as allergens or not.
How we built it
We trained the program to recognize ingredients when given a recipe. We then used a classifier trained on a dataset to predict whether the ingredients in the given recipe contain allergens or not.
Challenges we ran into
Finding extensive datasets was impossible which led to less accuracy than we would have liked Processing the recipe text to get the program to recognize ingredients Using machine learning algorithms to detect whether a given recipe contains allergens or not
Accomplishments that we're proud of
We integrated our NLP code with Google Cloud and Google Vision API (specifically text detection)
What we learned
Natural Language Processing Chrome Extensions
What's next for FoodAllert
FoodAllert will later include substitutions for allergens and will have a more efficient, accurate NLP algorithm so we can ensure that users get the best results. We also plan to have a feature where the user can specify what they are allergic to so that the results are more personalized.
Log in or sign up for Devpost to join the conversation.