Posted On: Oct 28, 2020

Amazon Textract is a machine learning service that enables you to automatically extract text and data, including from tables and forms within scanned documents and images. As a managed service, Textract delivers continuous improvements over time. Today, we are pleased to announce that we delivered enhancements to the underlying machine learning models resulting in a decrease of 20% in average latency when calling both the synchronous and asynchronous APIs. These improvements apply to the DetectDocumentText API and AnalyzeDocument APIs. 

As part of this update, we also improved the accuracy of detecting documents that are difficult to read because they are captured at extreme angles. 

The latest models launched today in all AWS regions where Amazon Textract is available. You will start noticing the decrease in latency and better accuracy of documents captured at extreme angles automatically without any action on your part.

Get started with Amazon Textract today.