Ripe

Overview

Summary

This iOS application was developed within ~21 hours for Major League Hacking 2017 and achieved 2nd place.

Ripe is a smart, all-in-one POS solution for medium-sized grocers. The Ripe iOS app combines an intuitive item select menu with neural net powered image recognition.

Motivation & Functionality

Many modern grocers and other retailers are switching from legacy, stationary checkout machines to flexible, mobile solutions.

Ripe allows a cashier to quickly build the customers order and take payment entirely through the app. The cashier can scan barcoded items with the optional Bluetooth scanner, select produce items using the visual selection menu, and in the case of hard to identify fruits or vegetables, use Ripe image recognition to obtain the item code.

Technologies

Neural Net

Used our own trained neural network running on Metal for the image recognition portion of the iOS application.

Swift Classes

• LoginViewController: Handles user authentication and landing page for app
• ProductListNavigation: Holds header and displays ProductListCollectionViewController
• CartTableNavigation: Holds header and displays CartTableViewController
• CartTableViewController: Shopping cart for selected items to buy
• CartTableViewCell: Item object with name, image, weight, and price
• SubViewController: Displays sub-categories for a chosen cagetory
• ProductListCollectionViewController: Categories for produce items
• ProductListCollectionViewCell: Category object with name and image
• AddToCartViewController: Modify weight or quantity of produce item
• CheckoutViewController: Handles user payment and generates a receipt
• SelectionViewController: Handles shared UI elements between ProductListNavigation and CartTableNavigation
• SelectionPageViewController: Controls screen interaction between ProductListNavigation and CartTableNavigation

What's Next

• Train neural net on large range of fruits, vegetables, etc.
• Provide Square integration for payment system during Checkout phase
• Add extended support for grocery item barcode scanning (via camera)
• Add extended support for weighing specific items (via Bluetooth scale)

Contributors & Past Experience

All members are currently 3rd year Computer Science students attending the University of Victoria.

• Ali Siddiqui: Moderate experience with Swift 3.x and Machine Learning.
• Matthew Paletta: Moderate experience with Swift 3.x and Machine Learning.
• William Grosset: Minimal experience with Swift 2.x.
• Jordan McKinney: Moderate experience with Machine Learning.