Smart Retail - Self Checkout Kiosk

AI Anti-Theft Sofware for Self Checkout Kiosk

Transforming Retail with AI: Minimize Theft with Cutting-Edge Technology.

Brief introduction about Smart Retail Solution :

Smart Retail is a Self-Checkout Kiosk systems that allow customers to scan and pay for their purchases without the assistance of a cashier. This service uses a self service kiosk or a dedicated checkout station, which is a combination of software and hardware designed to process transactions. The kiosk features a touchscreen interface, a barcode scanner, and a payment mechanism that supports a variety of payment methods.


Customers scan their products, bag them, and pay using a variety of payment methods, including cards and mobile payments. This approach allows for a faster, more efficient shopping experience while eliminating the need for staff at checkout counters making the check-out experience simple and convenient for all.


By integrating Smart Retail into your business, you will optimize operations while allowing your team to focus on critical tasks and on improving the rest of the services. Our solution lower the need for additional staff, allowing your employees to providing individualized and attentive customer care when needed. It also provides the highest ROI by increasing the number of daily transactions by taking care of more clients, lowering operational costs.


Smart Retail not only speeds up the checkout process, but it also increases transaction volume and lowers operational expenses enhancing your return on investment. In addition to these benefits, Smart Retail helps your business remain competitive in the evolving tech market where efficiency and convenience are key. It’s a solution that meets the modern customer’s desire for quick and hustle-free shopping while also supporting your business goals.

About the Customer

Let's Pizza Website :

Its success is based on the Al Munajem Group's over 60 years of knowledge, which has proven them as a trusted name in the business since their founding in 2014. With an strong focus on customer satisfaction, Laventa's machines use Smart Technology, which allows for real-time monitoring and quick refilling. Their commitment to innovation and quality guarantees that consumers receive only the greatest products, backed up by a dedicated staff that provides individualized service.


Laventa Vending Machines

Snacks Vending Machine Coffee Vending Machine Juice Vending Machine

Laventa has been a trusted partner for the past three years, recognizing the efficiency and cost-effectiveness of our Smart Retail Solution. These kiosks provide an affordable alternative than standard vending machines, with the added benefit of carrying more products. As a result, they started transitioning from vending machines to our Smart Retail solution, ensuring that they remain at the top of retail innovation while also expanding their service offerings.

Definition of the Problem

The fast growing use of self-checkout solutions in retail businesses has introduced a serious challenge; theft! Unlike typical checkout systems with a cashier being present at the counter, the self-checkout solutions rely on customers to scan and pay for their products. while these modern solutions have proven to be 50% more efficient than traditional ones, they also come with certain risks. Unfortunately, this reliance can be misused, as some customers intentionally fail to scan all their items, sneaking products through without paying. This form of theft is not uncommon and can cause significant financial losses for your business.


This problem isn't limited to isolated incidents. The lack of direct supervision at self-checkout kiosks can result in large revenue losses for businesses over time, especially in high-traffic areas with several machines. Laventa, a market leader in unmanned stores, has experienced a similar challenge. As they moved into self-checkout services, they discovered instances in which clients abused the system to avoid paying. This issue pointed out the need for stronger technology measures and improved awareness to prevent theft in current retail environments.


Addressing such issues is critical to sustaining profitability and operational efficiency. AI-powered systems can provide out of the box solution by monitoring customer behavior for any suspicious activities, thereby safeguarding the business from significant risks.

The proposed solution :

Overview of the solution:

To address the issue of self-checkout theft, we developed the AI Anti-Theft Module that not only minimizes theft but also provides real-time surveillance and records evidence of any fraudulent activities.


The AI Anti-Theft Module is a cutting-edge component of the Smart Retail solution, designed to enhance security and reduce theft in retail environments by leveraging artificial intelligence (AI) and machine learning technologies. This module integrates seamlessly with the existing infrastructure of retail stores, such as CCTV systems, to provide real-time monitoring and advanced threat detection capabilities.


Checking if the Customer is passing through the checkout kiosk. Detect if the customer is checking out all the products.

The key feature of this system is its ability to stream the CCTV camera directly to the NPU (Neural Processing Unit - Intel® NCS2), which is connected to our kiosk machine via a USB port. The AI Module will analyzes the footage in real-time, tracking customer behavior to ensure they are properly scanning and paying for all their items, ensuring a secure and reliable checkout experience.


Anti theft Checkout Architecture:

The following diagram provides a clarifying illustration on the system’s architecture:


Anti theft Checkout Architecture Diagram

In the process illustrated above, a customer enters the store and scans their products at the self-checkout kiosk. While the customer interacts with the kiosk, a CCTV camera records and streams their actions in real-time to our AI Anti-Theft Software. Our AI then transfers the camera data to our AI NPU USB Stick, where it processes the video data.


After processing, if any fraudulent action is detected, the system generates an alert. The event data, along with the complete recording of the scene, is sent to our dashboard. This ensures that the entire process is monitored and protected.

Hardware Specifications

Hardware informations:
  • CPU Rockchip RK3399 (64-bit CPU, 2.0GHz)
  • Android 10 operating system
  • RAM: 4GB ROM: 32GB
  • 24 “ Touch Screen: 10-point capacitive touch screen with 87±2% transparency.
  • 80mm Thermal Printer
  • 2D / 3D Scanner
  • Dimensions : 900mm x 450mm x 180mm
  • Floor Stand : with Stationery Table
  • Power Input: AC 100-240V
  • Operating Temperature: 0 ~ 40° (32 ~ 104F)

The Camera (CCTV or Internal):

Our Self Checkout Kiosk hardware can integrate an embedded camera, or an existing CCTV system in place already, we can directly connect its to our system using local video streaming protocol. The camera is essential for monitoring customer behavior during the checkout process.


Our AI Anti-Theft Software connect to the camera using RTSP (Real-Time Streaming Protocol) to read live video of the store. This enables the system to record and evaluate consumer activities in real time, ensuring that all items are scanned and paid for and providing critical data to detect and avoid any potential fraud.


AI NPU Intel® NCS2 (Intel® Neural Compute Stick 2):
A Plug and Play for AI Inferencing

The Intel® Neural Compute Stick 2 (NCS2) is a compact, low-power USB device designed for deploying and accelerating deep learning inference at the edge. The NCS2 is particularly useful for adding AI capabilities to existing systems like in our case self checkout kiosk scenario.


  • Processor: Intel Movidius Myriad X Vision Processing Unit (VPU)
  • Supported frameworks: TensorFlow*, Caffe*, Apache MXNet*, Open Neural Network Exchange (ONNX*), PyTorch*, and Paddle* via an ONNX conversion
  • Connectivity: USB 3.0 Type-A
  • Dimensions: 72.5 mm x 27 mm x 14 m
  • Operating temperature: 0° C to 40° C

Collecting the Data:

To build an effective AI Model for the checkout theft solution, we started by creating various theft scenarios and collecting real data using both the self-checkout system and CCTV cameras. This involved simulating different types of fraud, such as customers not scanning all items or bypassing the checkout process altogether. By capturing these scenarios, we accumulated a rich dataset of real-life behaviors, providing the dataset needed to develop and train effective AI model.


Training the models:

Using the obtained data, we trained our AI model on up to 1935 images, focusing on identifying theft patterns during the self-checkout process. The model has been refined to detect suspicious behavior and identify potential fraud. As a result of this rigorous training process, we achieved a remarkable 81% mAP (Mean Average Precision) accuracy, indicating that our system is highly effective at detecting and preventing checkout theft.


Results :

Empowering Retail with AI: Harness AI for Smarter Theft Prevention

In the demonstration video bellow, we showcase our AI-powered Anti-Theft solution in action at a real self-checkout kiosk. The video begins with a customer approaching the self-checkout, where they proceed to scan and purchase items.


The kiosk is equipped with a touch screen, a scanner, and a payment device for a seamless transaction experience, however, during the process, the customer accidentally forgets to scan a water bottle among the three items they intend to purchase.


Our advanced AI system, integrated with CCTV camera strategically positioned to monitor customer behavior, quickly detects the un-scanned item, the system identifies this as potential theft and immediately sends an alert to the operations team, ensuring immediate action is taken to address the issue. This demonstration highlights the efficiency and accuracy of our AI Anti-Theft solution in preventing loss and enhancing security at self-checkout stations.


Below, we have included images of the anti-theft application's interface. These images showcase key features such as the record button to initiate monitoring, the stop service button, the save button, and the configuration and setup screens. This additional visual content provides a deeper understanding of how the AI Anti-Theft solution works behind the scenes and highlights its ease of use.


Home Screen Area Selection Configurations Troubleshooting

Conclusion :

The AI Anti-Theft Software within the Smart Retail solution represents a significant advancement in retail security. By harnessing the power of AI and machine learning, this module provides real-time, intelligent surveillance that can detect and prevent theft more effectively than traditional methods. It not only enhances the security of the retail environment but also contributes to a better overall shopping experience for customers by ensuring a safe and secure atmosphere. As retail continues to evolve, the AI Anti-Theft Module positions Smart Retail as a leader in innovative, technology-driven solutions for the industry.

Contact Us

Smart Prints KSA

Address : 2930 northern ring rd - Al Wadi Dist. RIYADH 13313 - 6137

Phone : +966 552-385-990 / +966-543-694-064