Episode

What’s new in Computer Vision OCR

AI Show

Computer Vision just updated its models with industry-leading models built by Microsoft Research. These models are tagging contents in an image with significantly more detail & accuracy, across more languages. 

Jump to:
[00:16] Seth welcomes Sanjeev
[00:48] What is OCR?
[02:21] What's new to OCR? Multiple languages in same text line, handwritten and print, confidence thresholds and large documents!
[03:55] How Read operations works
[05:14] Demo - Read image as text
[07:05] Demo - Handwritten classification
[09:21] Demo - Mixed languages in same text selected pages extraction
[12:10] Seventy three languages supported by Read
[13:25] Demo - Text lines in Reading order
[15:49] Demo - OCR Applied (Form Recognizer) layouts and tables

Learn more:
What is Optical character recognition? https://aka.ms/AIShow/OpticalDoc
Quickstart: Read client library or REST API https://aka.ms/AIShow/RESTAPIQuickstart
How to call the Read API https://aka.ms/AIShow/ReadAPIDoc
Install Read OCR Docker containers from Computer Vision https://aka.ms/AIShow/OCRDockerDoc

Create a Free account (Azure) https://aka.ms/aishow-seth-azurefree
Get Started with Machine Learning https://aka.ms/AIShow/StartML
AI for Developers https://aka.ms/AIShow/AIforDevelopers
Azure Machine Learning https://aka.ms/AIShow/AzureML

Follow Seth https://twitter.com/sethjuarez
Follow Sanjeev https://twitter.com/SanjeevJagtap

Don't miss new episodes, subscribe to the AI Show https://aka.ms/AIShowsubscribe
Join us every other Friday, for an AI Show livestream on Learn TV https://aka.ms/LearnTV and YouTube https://aka.ms/AIShowLive