In the post yesterday I shared an example to show how to recognize text in pictures. The OCR API returns a collection of regions where the text is recognized.
If we go through each region can recognize points to create a frame, and inside it is the recognized text.
If we put together the facial recognition with text, we have a collection of Frames to add the image with different recognized elements.
Finally, an example with a meme and the great Ron Burgundy
With the following results
Text Angle: 0
Left 25, Top 8,
Height 396, Width 489
I’M NOT ANTI-SOCIAL
I’M SELECTIVELY SOCIAL
THERE’S A DIFFERENCE.
The source code is avilable in GitHub https://github.com/elbruno/ProjectOxford
Greetings @ Madrid !
- Project Oxford
- My Azure ML Vision APIs series
- My Azure ML Emotion APIs series
- My Azure ML Face APIs series