This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
You could follow the code step by step into OpenCV with a debugger if you build t with debug info (set BUILD_WITH_DEBUG_INFO in cmake) and configure your IDE to use them (see this how to with Visual Studio)
I would suggest to build and run performance tests from OpenCV. Enabling BUILD_PERF_TEST in cmake you will have a set of ready to use performance tests from OpenCV. Google C++ Testing Framework is used. See How To Use Perf Tests in QA in OpenCV
For example, to test MOG2 performance you could use opencv_perf_video. There is a perf_test for each module. Each test gives additional information about current parallel framework, ...
You could write your own test case using your data and same library/code (see How to Write Perf Tests) . On this way you will have an independent and repeatable performance test that might be shared with other user here.
In addiction you can test many functions at one time, this helps to understand where you gain and where you loose.
