Inspiration
I have always loved the google ai integration on android devices where when i hold my home screen button down, I can circle anything on my screen and get a visual search up about it, and even ask further questions about it.
What it does
So Cooking with Mia does/ is meant to do the same thing but for just food items/ meals. It's supposed to be an integration with other Ai's as a feature but for now it's a stand alone webpage that the user uploads the image of their food items and it searches the web for that food using Google's Vision API model and gets the ingredients needed and recipies for that meal and meals that are like that.
How we built it
I had a lot of help building it, I heard about it late and I was interested but my team were not prepared for it. I took used Google's Vision API model for the searching of food items from the images. I used Streamlit as the UI for displaying the gotten data. I wrote the entire thing in python.
Challenges we ran into
One of the challenges I could not figure out was actually getting the food images using the Vision API. I feel like the problem was that I could not come up with a training dataset to use for it but that's the major issue I had, other than that, everything works well.
Accomplishments that we're proud of
Creating the AI, and being able to have something close to what Google's circle and search can do really makes me happy. I am also proud of the fact that I have been able to learn a lot of things through my constant interactions with AI and my colleagues.
What we learned
How to make API calls, How to install dependencies, How to use Streamlit, how json files and API keys work.
What's next for Cooking with Mia
The next thing now is making it more accurate and able to predict more types of food and even bring out deeper information about the food and maybe, who can eat it and what health benefits you can get from it.
Built With
- googlevisionapi
- python
- streamlit
Log in or sign up for Devpost to join the conversation.