Skip to content

Latest commit

 

History

History
26 lines (19 loc) · 900 Bytes

File metadata and controls

26 lines (19 loc) · 900 Bytes

Python OCR image

This sample reads an image with Windows.Storage, decodes it with BitmapDecoder, and recognizes its text with Windows.Media.Ocr. It explicitly projects the returned content-type stream to IRandomAccessStream, so the required interface IID is validated before decoding.

The default run creates a local PNG containing DYNWINRT OCR 42, so the sample does not require a camera, network access, or external image asset.

.\generate.ps1 -Codegen ..\..\..\target\release\dynwinrt-codegen.exe
.\run.ps1 -Python C:\path\to\python.exe

Recognize another image:

.\run.ps1 -Python C:\path\to\python.exe -Image C:\path\to\image.png

Custom images are accepted without assuming their contents. Use -Expect WORD1,WORD2 to require specific recognized words.

Windows must have an OCR language matching one of the current user's profile languages.