C# using Project Oxford to convert images to entities via Vision API's OCR => LUIS
Microsoft's Project Oxford is a suite of MSR-built (with help from other internal teams) APIs for various high-level ML tasks, offered via the Azure Marketplace.
We focus here on their Vision API for image categorization and (in our case) OCR, and LUIS - their service for Language Understanding.
- Head to the API pages above and sign up for the services involved - LUIS is invite-only still so it might take some time to get in.
- Once you have the API keys, clone this repo (or fork/clone) and take a look at the code.
- Specifically, take a look at TestImageToText.cs and TestTextToEntitiesAndIntent.cs
- These tests are not proper "unit" tests since they invoke the services and don't mock them out - they're meant to make it easier for you to try things out. To use them, set the environment variables listed in the tests with the keys you've received upon signup.
- Note that for the LUIS-related tests, you'll need to have published a trained model. Follow the directions in LUIS to train and publish - it's easy.
Please take a look at my blog post on the subject for additional details, and hit me up either with comments on that or GitHub issues if you have other questions.
MIT