Add a new variant of detectMultiScale with output arguments 'objects' and 'numDetections'; factor most of detectMultiScale's code into 2 protected methods

It was merged by mistake.

This reverts commit ab6be9b7.

receives the number of neighbors joined into each detected object

This should allow many algorithms to take advantage of more parallelization
technologies.

Also move cv::linemod to own header

This also fixes few clang build errors

cv::Complex, cv::Point_ and cv::Point3_ are moved.

Also move RTTIImpl class to the only usage

e.g. <opencv2/core/core.hpp> become <opencv2/core.hpp>

Also renamed <opencv2/core/opengl_interop.hpp> to <opencv2/core/opengl.hpp>

Made all STL usages explicit to be able automatically find all usages of
particular class or function.

Declare vars as locally as possible

disabled the use of SSE4 instructions as well to make the code compatible with the old Intel and AMD chips

added test for the old cvHaarDetectObjects. disabled AVX optimization in haar.cpp. it should cover tickets #2534, #2591, #2669

 - (left, top, width, height)
 - (left, top, right, bottom)