How does an AI-based OCR technology work for automatic codes reading?

The disruption brought by the incorporation of AllRead’s Artificial Intelligence in supply chains and Industry 4.0 or water and gas companies is invaluable. AllRead can detect and digitize any text or code from any IP camera or mobile device, by installing the on-premise or SDK solution. Discover how it works, the added value it offers and its integration possibilities.

We’re pioneers in using Machine Learning technology to offer a state-of-the-art OCR software, with which companies can automate their processes. AllRead develops a technology based on deep learning to detect and digitise text, codes and symbols. It helps to automate data capture processes, reduce repetitive manual tasks and improve operational efficiency.

Our Automatic Reading Software (ARS) automatically processes images taken by a mobile or fixed camera and provides digital extraction of the information of interest. ARS is integrated as an SDK in an APP, or installed “On premise” on the customer’s server, connected to IP cameras.

It detects, reads and digitises serial numbers, sensors or meter readings, barcodes, license plates and many more structured codes in complex operating environments. Then, it sends and integrates the data into the existing information system.

AllRead presents a deep learning technology to detect and digitise text, codes and symbols in supply chains, Industry 4.0 and utility companies from video and images.

Adriaan Landman, COO and Cofounder of AllRead
AllRead’s ARS software reading utility meters consumption numbers from a mobile device.

How does this technology improve on other OCR solutions?

AllRead is not another OCR, but an information detection solution, adapted to each specific problem with fast training of the learning algorithm. AllRead achieves automatic and accurate extraction of text, codes or symbols where other traditional OCRs fail. It resists motion, blur, wear, rotation, smudging… which are common challenges in operational environments.

With the ability to read serial numbers, sensor or counter readings, ID codes and other process-standardised data formats, intelligent text recognition represents a new way to collect data in supply chains.

The competitive advantage of AllRead is that it combines:

  1. The highest accuracy in data extraction, along with
  2. Great agility for implementation, adapting to existing processes and hardware
  3. Avoids obsolescence by being able to develop new reading fields and retrain to new casuistries
  4. Thus, offers a better return on investment compared to existing solutions.
Utility meter
Example of meter that AllRead’s SDK solution can read from any mobile device, even when photos are blurry
or codes are damaged or partially hidden.

Which customers use and benefit from this technology?

We assist seaports and port terminals, such as the Barcelona Port Authority and Hutchison BEST Terminal, with the traceability of assets transiting through their facilities (containers, wagons, trucks, dangerous goods…) through the detection by fixed cameras of their identification codes.

In this way, we speed up incoming and outgoing goods, reduce bottlenecks and the associated CO2 emissions. As for gas and water distributors, such as Naturgy or Aguas de Valencia, we offer them a mobile reading solution for the meter data collection: consumption and identification numbers.

It enables a better user experience —be it for the end consumer or the employee in the field— and optimises information management processes for the company.

AllRead operates mainly in two verticals: port logistics (asset traceability) and water and gas supply (taking meter readings).

Can your technology be incorporated into other products or integrated into an App?

Our software is delivered in the form of a REST API and SDK, which allows our customers to easily integrate it with their existing information systems, or in their APP.

Our software can be integrated via API Rest and SDK.

As a startup, what challenges have you encountered and what are your projections?

In addition to the usual challenges that startups face, such as finding their target market, hiring and retaining talent, developing the product… we face long sales cycles, due to the type of solution we sell (industrial software) and the size of our customers and prospects.

So, although our value proposition is received with a lot of enthusiasm, there are several steps to manage before we get to the actual implementation. The company was founded in 2019, and today we are about 20 people involved in the company.

Our goal is to consolidate our position in the mentioned current market niches, and to extend our reach to other segments of Smart Logistics.

Which is the distinctive value of AllRead’s AI-based OCR software?

There are many OCR solution providers, especially giants such as Amazon Textract or Google Tesseract, which offer generic OCR at very competitive prices or Open Source. In itself, OCR no longer represents any competitive advantage.

Therefore, the important thing in my opinion would be specialisation in a market segment with specialised OCR solutions, and/or the development of an own OCR, with unique features including AI of course. At AllRead, we have relied on these two differentiators.

Original source: Entrevista a Adriaan Landman, Co-Founder & COO de la startup AllRead, published on 4th November 2021, by Doonamis:

Leave a Reply

Your email address will not be published. Required fields are marked *

2022_AllRead Logo_Negativo

Norrsken House Barcelona. Passeig del Mare Nostrum, 15 08039 Barcelona, España

© 2024 · AllRead Machine Learning Technologies S.L. All Rights Reserved.