Algorithmic decision making
Self-learning and adaptive.
By analyzing millions of images of shopping carts and baskets, Cartwatch learns the key features associated with these objects and builds an abstract representation which can be applied in diverse retail environments.
In the retail sector, open entrances have been found to have a positive effect on the shopping behavior of typical customers. However, it also enables thieves to steal fully-loaded shopping carts or baskets in a process known as pushout..
At Cartwatch, we understand this problem and have developed a minimally-intrusive solution to eliminate this mode of stock loss.
Using state of the art methods from the fields of artificial intelligence and computer vision, we are able to detect and track full carts and baskets.
In the event of prohibited motion, our system is able to trigger alarms and capture video evidence of the event, thus making theft increasingly difficult, dangerous and therefore unattractive.
Our detection technology can not only be used to protect open entrances but also for instance as an integration into other loss prevention systems at the POS.
At the checkout, customers will often, intentionally or otherwise, fail to place all of their items on to the conveyor belt for purchase. Small, high-value items are particularly prone to being left in the cart, leading to shrinkage for the retailer.
If the cashier notices the event occur, apologizing or feigning ignorance will usually suffice to prevent further questioning. As a result, this tactic is essentially risk-free. This makes it both one of the most common methods currently used by regular shoplifters, and an introduction to shoplifting for otherwise honest customers.
Due to the ubiquity of this method, it is a major contributor to shrinkage. Standard countermeasures, such as installing angled mirrors or having the cashier stand to check each cart, implicate honest customers and slow down an otherwise highly-efficient checkout process. This reduces revenue and increases cost for the retailer.
Cartwatch Checkout is an AI assistant for the cashier. Using a camera, it screens all shopping carts passing through the checkout. It analyses images of a given cart to determine if items are still present, guides the cashier’s attention and enables them to review suspicious events on the checkout computer display.
- Reduces stock loss at the checkout
- Improves the checkout experience and shopper throughput
- Improves cashier’s focus on real, suspicious events
- Is invisible to the customer - no false implications
- Tests show that once detected, hidden goods are usually paid for
Employing only a small system of cameras, Cartwatch replaces physical entrance barriers with a virtual barrier, depicted as a red line in the demonstration video.
The placement of this barrier can be individually tailored to each store during installation. Cartwatch then autonomously monitors the entrance area of the store and automatically detects prohibited motion of full carts and baskets. In the event of attempted theft, various kinds of alarms can be triggered.
Security and data privacy
Customer privacy and reliable differentiation between permitted and prohibited motion are at the core of our product. Cartwatch operates on anonymized data, satisfying data protection requirements, and monitors only full carts or baskets.
This leaves honest customers with empty carts and baskets, or those moving in permitted directions, with the freedom to enjoy the retail experience they expect.
Cartwatch bietet vom 6.-7. September 2018 ein Seminar für Künstliche Intelligenz für Entscheider im Handel an. Registrieren Sie sich bei Interesse bitte hier.