The problem SeeFood solves

Do you find your favourite fruit in your refrigerator basket at the time when your mood is so much for that or are you frustated of daily answering to your mom what to make in lunch/dinner and sometimes you need your favourite meal and diidn't get it?
Then we have SEEFOOD, an app for such foody persons who love to eat fruits daily, different variety of food items in their daily meals. It monitors the fruit you store in the basket and the times you take it out along with vegetables and if it happens that any items finishes, it sends a notification on the mobile of particular person. It will also act as an assisstant to your mom since it happens often, she gets confused what to make with the available website. Searching on youtube takes a little longer time. Let us make that easy for you, SEEFOOD also provides assisstance to what you can make of the available food. With SEEFOOD, keep a track of your diet and your family members diet depending upon what they take out from basket.

To Sum Up,

SeeFood helps you to monitor fruits, vegetables and other eatable items stored in refrigerator. If these items finish it will let you know so that you can buy new one before getting frustated and what you can make of the available food at your home by providing different recipies, also it will help you to keep a track of your diet by sensing the fruits/vegetables that you take out of the refrigerator.

Challenges we ran into

There were lots & lots of errors that occured while we were trying to train & run our Machine Learning Models. All of us, just put all of our brains scrolling and reading through all the relevant documentation we could find on tensorflow, stackoverflow, reddit etc.

Then, when all the errors were gone, our ML model worked with unbeliveable accuracy! That was the happiest moment of our team during this whole hackathon.

But, later on at one point of time, our ML model started to give huge inacurrate results, like for image of Banana, it predicted it was an apple and vice-versa. So, we put a lot of time again into training that model and then converted it into keras.