I accidentally bought the same terrible frozen pizza brand twice 😖 and vowed never to let that happen again to me or anyone else! 🥶🍕🙌
PizzaVision is a cross-platform mobile app where you can snap pictures of product packages and tag them as either "👍" or "👎".
Forget if you've tried a product before? Just snap a pic of the packaging, and the app will search your photos and show similar ones you've had in the past.
Download PizzaVision on the App Store
Live Demo PizzaVision Web App
(Android version to come)
With so many products in stores, frozen pizza stands out as one of the most cluttered category. They all look alike with Italian-sounding names and picturesque images, but few are worth their price.
How do we remember which ones we've liked and which to avoid?
Enter PizzaVision: snap a pic of the package, tag it "👍" or "👎", and never forget your preferences again.
As my final project for the Harvard CS50x course, I wanted to learn new technologies and hone my product developement skills, while tackling a meaningfull problem for society.
PizzaVision allowed me to learn new technologies, particularly mobile development and computer vision, by tackling a common problem: remembering product preferences.
While I had postpone the broader societal impact bits, I certainly solved my own problem!
git clonecd pizza-vision\ionic-appnpm i
You don't need to sign up for a Supabase account to run this code locally. Just run Supabase locally on your machine.
- Install the Supabase CLI, along with Docker.
- Skip
supabase initas the supabase folder is included. supabase start- add a .env file to the ionic-app folder with the following:
REACT_APP_SUPABASE_URL=http://localhost:54321
REACT_APP_SUPABASE_ANON_KEY=(the anon key displayed when you start Supabase)
- Open Supabase Studio on your localhost, go to "Storage" and add a bucket called "photos".
This project doesn't allow direct collaboration on backend code yet, but you can deploy to your own Supabase project.
More details here
npm run startfrom the ionic-app directory
- Ionic: mature framework with large community support.
- Capacitor: cross-platform runtime for accessing mobile features.
- React: popular library well integrated with Ionic.
- @tensorflow/tfjs with @tensorflow-models/mobilenet.
- Supabase: relational database, Auth, object storage, basic CRUD API, realtime API.
- User authentication via one-time password with email.
- Basic user profile management.
- Snap pictures with the phone's camera.
- Upload directly from the camera or from the phone's photo library.
- Tag pictures as "👍" or "👎".
- View similar images ordered by degree of similarity.
- Library view ordered by date taken.
- Filter library by tags.
- Sync local images with cloud.
- Improve performance of image processing.
- Implement infinite scroll on home feed.
- Add account management options.
- Image optimization.
- Library sharing.
- Automatically classify images using EfficientNet (predictions from MobileNet aren't accurate enough for our use case).
- Filter and reorder library view.