'How are openCV matrices serialized in Google's mediapipe
Google's mediapipe framework is able to work with openCV matrices across different languages, certainly C++ and Python.
Since it uses protocol buffers to communicate data among (computational nodes) I'm trying to find out how "image messages" are created. So far I can identify these related .proto definitions and sources:
- image_format.proto
- image_file_properties.proto
- image_frame_opencv.h (and respective implementation file)
- image_frame_util.h
But I don't get the whole picture. Where is an "image" defined the mediapipe protocol buffers? And are there helper classes per language, most importantly C++ and Python?
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|