- Recognizes the container manufacturer with the user pointing his or her phone camera onto the logo.
- Represents information on plastic type and recyclability.
- Identifies the user location automatically.
- Lays out the route to the nearest recycling point.
- Ensures QR-code scanning to get Helpful-tokens.
- Enables token exchange for eco-friendly goods.
Client and Business Goal
The aim of a London startup was to motivate people to lead a plastic-free life and switch to environmentally recyclable bottles. The client ideated an application that would reward people who abandon disposable plastic containers with in-app currency to be later exchanged for eco-friendly goods.
The idea implementation required cross-platform mobile and web development skills, as well as machine learning competency for object recognition. Thanks to our proven expertise in AI, ML, and digital currency technologies, AIS Novations was selected as the software engineering partner for the project.
We used Apache Cordova tools to develop a hybrid app, and Azure cloud platform to ensure scalability. The CMS was implemented as an SPA and the web app was developed using a classic approach with server-side rendering. The application allows scanning the label on the plastic tare and/or tare object recognition in real time and suggests the location of the nearest recycling point.
We used Azure Custom Vision Service for visual analysis and image recognition. To enable a live search of the object in the frame, we applied TensorFlow models that launch on any user’s device and operate faster than on the API. TensorFlow, a machine learning library by Google, allows building and training neural networks to automatically locate and classify images, thereby achieving the quality of human perception.
Several Cordova plugins demonstrated unexpected behavior and lacked some of the required functionality. We decided to debug and customize them, both for iOS and Android platforms.
The project was completed successfully despite the tight timelines and budget constraints. The hybrid app powered by the latest AI/ML tech for visual object recognition helps identify and recycle plastic tare of more than 20 brands, including Coca-Cola, Pepsi, Sprite, Tropicana, Starbucks, and others.
Over 15 thousand units of disposable plastic bottles and cups were recycled in the first few weeks after the startup launch. The project is expecting to attract investment for their initiative on bathing water and beach clean-up.