fruit image recognition
The big variety of fruits at Hack Zurich inspired us to motivate others to eat more fruits. We wanna do this in a fun and easy way. We were inspired by the recent development in the field of ML and wanted to use some of those technologies for our project.
What it does
It shows you challenges you can accept. For example: "You have one week to eat an apple, a pineapple, a banana, a pear and an orange. Each day you choose one of these fruits, take a photo of them with our App and eat it afterwards." AppleOrDie uses AI to verify that it's the right fruit and keeps track of your progress for you. You complete a challenge if you manage to eat all the fruits in one week.
How we built it
We use the Google vision API to recognize the different fruits and reject wrong ones. We preprocess the image data we get from the smartphone camera to greatly enhance the speed of the recognition process.
Challenges we ran into
Debugging React Native on a iOS 11 beta device can be quite hard and frustrating. It really helps to develop for stable software.
Accomplishments that we're proud of
We are proud of our teamwork and that we built a fun to use app that really motivates to eat more fruits.
What we learned
First read documentation then start implementing.
What's next for AppleOrDie
We want to build on top of our existing idea by making it easier to compete against friends and family. The process of snapping the fruits has to be an extremely pleasant process. We could imagine to add a reward system where points can be traded to goodies like fitness watches or result in a small deductible on your health insurance premium. Also food producers could develop their own challenges and encourage the users to try their newest kind of apple.