DEV Community

Iren Korkishko
Iren Korkishko

Posted on

Computer vision for cashierless checkout

Hey fellows! I want to hear your thoughts about the Cashierless Checkout we've made. This solution based on object recognition and aimed at eliminating the need for traditional manned checkout counters, allowing customers to enjoy their shopping experience.

Image description

For this Cashielress Checkout we utilized computer vision algorithms to accurately identify the items chosen by customers, calculate the total cost, and securely process payments. The software processes video stream in real-time, so after the product item crosses the counting line its quantity in the shopping bill is incremented. Accuracy we have reached is about 97%.

https://www.youtube.com/watch?v=lEWFmKI5RYY&t=1s – here is the example of how it works.

When working on item recognition, we took into account:

  1. accommodating various product sizes,
  2. shapes, and
  3. packaging types across different retail sectors.
  4. Also, product overlapping.

By implementing Cashierless Checkout, retailers can optimize operations, reduce staffing costs, and redirect resources to other critical areas.

Image description

Your opinion needed of what else we can take into account to improve the solution. Thanks!

Top comments (0)