Eat Safe Eat Well!

Input Banana

Inspiration

Have you ever heard of sayings and warnings about food contradictions like "don't eat carbs and persimmons together"? Overwhelmed and concerned about your food choices now? Inspired by these "wisedom from the past" and scientific studies, our program attempts to provide instant information regarding these food contradictions.

What it does

Given a photo of a fruit or a vegetable, our program will list out all the food contraditions found in the database.

How we built it

Through pre-trained CNN we classify input image into specific types of fruit or vegetable. Given the labels of the item, we search for rows in the contradiction table for matching item, and output an image demostrating the result.

Challenges we ran into

Image classfication has been wildly done, and looking for a topic to work on was difficult. We also had trouble sharing dataset and training out own model due to the time constraints. Similarly, we reduce the number of food categories significantly to save time in training and testing our program.

Accomplishments that we're proud of

What we learned

Brainstorming and beginning a project. Version control and sharing codes through github while working simultaneously.

What's next for Eat Safe Eat Well

Use more data of images for more types of raw food ingredients; collect more systematic dataset on the contradictions (as currently all contradictions are from unrelyable online sources). Better result presentation than an image. More complete UX/pacakge program into a mini mobile app.