Company
🔥 Just Launched
Introducing the AI Playground — Your LLM Battleground to Test Powerful AI Models!
Try now
Contact us
Contact us
Join the Discord
Join the Discord
December 10, 2020

How to Use Clarifai's Text Recognition Workflows

Clarifai Computer Vision By Jeff Toffoli

Table of Contents:

Read and Understand Text in Images

We are retiring our OCR model and have replaced it with a new and improved Visual Text Recognition (VTR) workflow. The new workflow separates the components of text recognition so that you can work with detection, cropping and text recognition as individual components in your projects. 

 

Join our Slack Community.

Join our community of users! Post questions, follow discussions and share knowledge.

Join Slack Group

icon-slack-community

The VTR workflow is already configured for you when creating an app in Portal. Just select “Visual-Text-Recognition” when selecting the base workflow for your app. 

vtr-base-workflow

You can inspect the individual models in the workflow by clicking the “Copy to new Workflow” button in the “App Workflows” section of the App overview page.

visual-text-reco

The VTR Workflow Post Workflow Results Request in our API 

 

The new VTR workflow uses a call to the workflow endpoint in our API (as opposed to the OCR model where you were making a single “model predict” call).

 

curl --location --request POST '{{base_url}}/v2/users/{{user_id}}/apps/{{app_id}}/workflows/Visual-Text-Recognition/results' \
--header 'Content-Type: application/json' \
--header 'Authorization: Key {{key}}' \
--data-raw '{
"inputs": [
{
"data": {
"image": {
"url": "{{your_image}}"
}
}
}
]
}'
view raw vtr_workflow.curl hosted with ❤ by GitHub

 

Click here to find out more about making workflow predict calls.

Previous Return to Blog Menu Next