Summary
How to correctly execute preprocessing in OpenVINO™ C++ API inference.
Description
- Created a custom C++ and Python* code for OpenVINO™ inferencing.
- Used the same image and models for both inferencing processes.
- Obtained different results from the C++ code as compared to Python* code.
Resolution
- Incorrect results were due to incorrect filling of tensor data in the preprocessing part of the C++ code.
- Ensure the channels when preprocessing the image matches the channels of the model.