Inspiration
Image2Plate was inspired by the need to bring people closer to the art of cooking and create a global community of food enthusiasts. We wanted to leverage technology to simplify the process of discovering new recipes and cooking inspiration, regardless of one's culinary expertise.
What it does
Image2Plate employs computer vision (ResNet101 model) to recognize food items in user-uploaded images, then utilizes a powerful language model to generate recipes based on those ingredients. Users can access nutritional information, share their culinary creations, and connect with a diverse community of food lovers from around the world.
How we built it
Image2Plate was built with a variety of technologies. We used Streamlit as our web framework for both the frontend and backend. Google Cloud's Firebase was used to make user authentication possible on our Streamlit app. We selected the HuggingFace ResNet101 model API to recognize the foods in the images inputted by the user. We used the OpenAI API and prompt engineering to generate the recipes and nutritional information of the recipes.
Challenges we ran into
Some challenges we ran into included testing and comparing various computer vision models to classify the food images inputted by the user as well as building the various features of the application including the social media timeline/feed and user authentication.
Accomplishments that we're proud of
We are proud to have made a fully working user-friendly website that takes in images as input and classifies the objects in the images to notify the users of possible recipes that could be made as well as the nutritional information. We are also happy to have added a social media component to this website.
What we learned
We learned and gained experience working with Streamlit, GCP Firebase, machine learning, and APIs.
What's next for Image2Plate
- Recreate Image2Plate with a web framework that can scale to more users compared to Streamlit.
- Perform customer discovery research by continuously interviewing users and potential users to get constant feedback on ways our platform can improve.
- Create a mobile app for both Android and IOS.
- Explore and implement various monetization models such as premium content and ads to support the platform's growth.
Log in or sign up for Devpost to join the conversation.