The OCR is Dead, Long Live Textract!

OCR technology

(3 minutes reading)

Amazon Textract is receiving significant hype in the OCR landscape. Since its launch in November 2018, countless articles and press releases have been published discussing the Textract’s potential to revolutionize the OCR industry.

Now that the dust has begun to settle down, it’s time to look back and understand to what extent this potential has been realized and what the real impacts of Amazon Textract were on incumbent OCR technologies

What is Amazon Textract?

According to Amazon, Textract is a service designed to extract text and data from virtually any document.

Textract combines proprietary OCR technologies with machine learning, taking advantage of the same deep learning technology that is currently used by Amazon’s computer vision scientists to analyze billions of images and videos daily. This approach ensures continuous learning through new data and allows Amazon to deploy new features continuously.

On the business side, Amazon Textract has a very competitive price that starts at .00015 cents per page for basic OCR or $1.50 for 1,000 pages.

Amazon Textract is Amazon’s attempt to move to the uncharted waters of OCR 2.0 and establish itself as one of the pioneers in this space.

Is Textract a Game Changer?

Leaving behind all the buzz and marketing cliches, the reality is that Textract seems a bit dated based on what we’ve seen in recent years, considering both its breadth of features and uses cases.

Amazon fails to capture the capabilities of modern commercial OCR technologies since many of the challenges they point out as limitations in these solutions have been solved for quite some years.

It’s also worth mentioning that, based on what Amazon presented in their launching webinar, Textract is quite behind in their detection, understanding, and processing of table data. Despite that fact, this was presented as a unique selling point of Textract.

Despite these limitations, we view Textract as a promising opportunity to push forward OCR technologies and, most of all, change and educate the market on new trends that are not being addressed by incumbent technologies.

Textract is not closing any doors to OCR solutions. It is, in fact, laying the groundwork for the development of new and improved data capture solutions.

Textract’s competitive edge against low-level OCR providers will be in using Amazon’s scale and access to data to pressure them on price.

Textract

Amazon is not alone in this game; Google Vision and Azure Cognitive Services are offering roughly the same features with a similar market approach.

Read more about Textract here.

What is DocDigitizer?

DocDigitizer is a Data Capture service that leverages both proprietary and third-party technologies.

Our value proposition goes way beyond offering image to text recognition or providing unvalidated generic data structures.

As digital document inbounds grow in quantity and diversity, automation must have the ability to understand semantic information rather than following hard-coded rules or configurations.

There was a time where automation meant processing scanned structured documents in back-office operations.

Nowadays, businesses are not able to impose strict formats on the way they communicate with their customers. Digital transformation simultaneously offers greater diversity, more communication channels, and faster response times.

Robot Process Automation (RPA) is also shaping the automation landscape with a tremendous impact on data entry related processes. Smart, scalable, and precise data capture solutions are key to supporting both Digital Transformation and RPA initiatives.

At DocDigitizer, we are democratizing the access to cognitive data capture, providing the following benefits over more traditional approaches:

  • We process any document (structured or unstructured) independent of its language, layout or domain.
  • We don’t have any setup cost.
  • We provide fully validated data with 99% accuracy guaranteed.
  • We leverage state of the art proprietary Machine Learning and Natural Language Processing technologies.
  • We combine the speed and scale of machine data capture with the assurance that can come only from our crowdsourcing of data curators.

All this is included in the form of an API as a service, providing the automation enabler that you need to support any digital transformation or RPA (robot process automation) initiative.

Get rid of your legacy OCR technology and upgrade your automation initiatives!