@dvd42 I think your model if failing builds because of "std::to_string()" instead use string stream do something like

stringstream ss ;ss << prediction.first;string pred = "Prediction " + ss.str();

and similarly for confidence.
Do it for all the files
This is a known problem of Android GCC toolchain + gnustl_static C++ runtime.

Please extract code from text documents and embed it through@include or@snippet doxygen commands.

Dedicated locations for tutorial files (or reuse any existed sample from "samples/" folder):

• https://github.com/opencv/opencv/tree/master/samples/cpp/tutorial_code
• https://github.com/opencv/opencv/tree/master/samples/python/tutorial_code

This help to ensure that code in documentation is not broken:

• All C++ samples are checked through compilation on CI.
• Python files are checked by linters (pylint / flake8 - see "Docs" builder).

Hi,@alalek, thanks for the feedback. I have a couple of questions:

1. Will the CI be able to check the samples if they have runtime parameters such as the ones in detection.cpp or classification.cpp?

2. The code that I have inserted in the markdown is simpler than the one in the samples, since I wanted to make the example easier to understand, can I include a block of code with@code ...@Endcode instead of embedding a file? If not, where should I create a new file and put this block of code in order to embed it using@snippet?

3. Some of the required files for the samples to run are network weights, these files are too large to make them part of the tutorial files and upload them to the repo, what should I do?

1. CI just checks compilation (in case of C++). Samples are not launched.

2. Normal "samples" contains a lot of helper code, like args parsing, etc. Code in tutorials is more simple. There are examples ofcode block andits embedding

3. It is better to reuse already existed data in repository. If not, then external link (URL) is preferable in case of huge data. Users should be able to reproduce tutorial results.

3. The external data that I use its already part of the downloadable links inopencv_extra, should I just explain this in the tutorial, or just comment the URL in the tutorial file so that the users can download it?

I believe, it is enough to add reference with model's name and with the link on downloader script.

This method should be called once

Please modify Python sample as well.

Please replace to

int classId;float confidence;std::tie(classId, confidence) = net.classify(frame);

In fact, we can just push new colors in case of new class id is more than current number of classes. So please modify code so it can manage it.

Hey,@dkurt I am sorry, but I dont understand what you mean by this, could you please explain with an example. Thanks.

unused header