Next-generation automated checks powered by edge AI to transform the accuracy of inventory management.

In recent years, there has been a strong demand for more sophisticated inventory management in the logistics, retail, and manufacturing industries.

Amidst labor shortages and increasing reliance on individual employees, preventing inventory discrepancies and shipping errors has become a crucial issue that impacts the quality of on-site operations.

Against this backdrop, barcode management has been introduced in many workplaces, making a significant contribution to improving inventory accuracy.

On the other hand, less than 25% of companies are able to maintain an inventory accuracy of 80% or higher (*), and the reality is that there are many cases where discrepancies occur between system-based inventory and actual inventory.

*Source: Food Logistics “Robotics Ranks No. 1 Solution to Improve Inventory Accuracy: Study

Many of these factors stem from human error, such as missed scans, misscans, and data entry errors.

While the barcode reading accuracy itself is extremely high, the question of "what, when, and how to scan" still depends on human judgment, leaving operational challenges.

In other words, current inventory management is a combination of "high-precision tools" and "manual operation," and has not yet reached complete automation.

In this context, a new approach to inventory management that utilizes AI, specifically OCR (Optical Character Recognition) and image recognition, to automatically recognize characters and codes from camera footage is attracting attention.

This article, accompanied by a video, showcases how edge AI can be used to streamline inventory management operations by leveraging OCR and barcode recognition.

Case study: Automating inventory checks using OCR (optical character recognition) and barcode recognition.

In warehouses and store backrooms, checking label information is a daily practice when receiving, shipping, or replenishing goods.

Checking expiration dates and best-before dates is especially important for pharmaceuticals and food products, as incorrect handling can lead to significant risks.

In this demo, an AI running on an edge device simultaneously performs OCR (optical character recognition) and barcode recognition based on label information acquired by a camera, and makes a decision on the spot.

For example, when an employee scans a product label, the system automatically determines the expiration date and displays it in green on the screen if there are no problems.

If an item has expired, a warning is displayed in red, making it an intuitive UI to understand.

This eliminates the need for manual visual inspection, enabling speedy and accurate inventory checks.

Here is a demo of inventory checking using OCR and barcode recognition (video in English).

In this demo, the following processes are performed on the edge device:

① Extraction of label text information using OCR

② Acquisition of product information by barcode recognition

③ Expiration date determination based on business logic

④All processing is completed within the local environment

This allows for real-time decision-making on the spot, without relying on the cloud.

In addition to stable operation unaffected by communication environment, a major advantage is that there is no need to transmit confidential information externally.

Another notable feature is that it utilizes image recognition technologies such as OpenCV 's Paddle OCR (character recognition) and Zcrossing (barcode recognition), and executes these processes on a single SoC.

By integrating systems that previously consisted of multiple devices, the burden of implementation and operation is significantly reduced.

Features of edge AI that combine flexibility and ease of deployment

One of the key features of this system is its high degree of flexibility, thanks to its template-based configuration.

When adapting to new labels or product variations, it is not necessary to retrain the model as in the past.

This can be done simply by specifying an area of interest (ROI) and defining the items you want to detect as a template.

This setup is completed quickly and will then operate stably on a continuous basis.

​This significantly reduces the adjustment work and line stoppage risks that occurred with each product changeover.

These features give this solution the following advantages over conventional machine vision:

• Expensive specialized equipment is not required (low cost).
• A simple configuration that is self-contained within a single SoC.
• Flexible and adaptable, no need for retraining.
• Stable recognition performance independent of rotation angle, lighting, and distance.
• Global expansion with support for 80 languages

In particular, its ability to be less affected by environmental conditions and to adapt to on-site conditions in a short time significantly lowers the barrier to adoption.

Summary: Inventory management is shifting from a "human-based verification process" to a "system where AI makes decisions."

The example presented here demonstrates how combining OCR and barcode recognition can eliminate issues such as missed scans and human errors due to reliance on visual inspection, enabling automated inventory management on the spot.

Furthermore, its flexible, template-based configuration allows for quick adaptation to product changes, and its simple architecture, consisting of a single SoC, makes it low-cost and easy to deploy—both major advantages.

​AI that operates entirely at the edge is evolving beyond being merely a tool for improving operational efficiency; it is becoming a foundation that automates on-site decision-making and raises the overall quality of work itself.

​The SiMa.ai​ ​MLSoC™ used in this demonstration is a next-generation chip optimized for realizing AI that can be used in the field.

Despite its compact size and energy-saving design, it achieves highly efficient inference of up to 50 TOPS and supports high-speed production lines with real-time processing at 120 FPS. The flexible development environment provided by the built-in Arm Cortex-A65 is also a major benefit of its introduction.

If you are considering introducing AI, you can start small like in this case study to verify the improvement effects on your own production line.

We hope you will find this article useful as your first step in utilizing AI.

Inquiry

Please feel free to contact us with any questions about our products, technical inquiries, sample requests, or estimates.