Artificial Intelligence enabled Cashier-less Retailing
AI enabled Cashier-less shopping was demonstrated by CellStrat AI Lab members – Abdul Azeez, Bhavesh Laddagiri & Shreeyash Pawar, through their product SenseCart.
How many times have you gone to your favorite retail store and didn’t make a purchase because of the long checkout queue? As a store owner, aren’t you bothered by inefficiencies in manual billing, absorbing time that customer could’ve used to purchase more items? Shopping carts have been around for ages, with no innovation, we at SenseCart, integrate AI with heterogeneous computing to provide seamless shopping experience, keeping customer satisfaction as our prime focus. Using AI, we recognize the products being placed in the cart and create a cashier-less, customer-centric and hassle-free shopping. This results in increased footfalls, lower resources, and increased efficiency. Now, who would want to return to a store that has the same old cart again? Digitize with SenseCart.
SenseCart Recognizes products captured by the On-Cart Module by using a Siamese Network which semantically compares its feature vector with the feature vectors of the product dataset and accordingly ranks the products from the dataset based on similarity.
The On-Cart Module consists of two Infrared Sensors and a camera. The sensors help understand if the product is being added or removed from the cart. When the sensors get triggered, it captures an image and POSTs it to the Inference API.
The Inference API is hosted on Google Cloud Functions which receives the Image and uses their model for semantic comparison of the query image with the images in the product set. The result are then updated in Firebase Real-time Database.
Firebase Realtime Database is where the User’s Cart and Sessions are stored and also acts as the serverless backend for SenseCart’s Web App.
SenseCart’s React App allows the customer to start a cashier-less shopping session with just a QR code scan on the cart and keep track of the cart and make payment.
View complete Poster here.