autorotate in-memory jpegs (#8492)

* autorotate in-memory jpegs

* correct indentation (4 spaces)

* imgcodecs: don't apply EXIF rotation for unloaded images

* videoio: don't try to rotate MJPEG stream

* imgcodecs: ByteStreamBuffer::seekoff support all seek "dir"

* imgcodecs: fix condition: "off == egptr() - eback()" is valid offset

Add method KeypointBasedMotionEstimator::estimate(InputArray, InputArray) to support both cpu & opencl algorithm processing

the orignal estimate function has input parameters defined as "Mat", this prevent users to call into algorithm opencl path

The method does cvCheckPixelBackgroundNP (which reads bgmodel)
before it is ever updated (in cvUpdatePixelBackgroundNP). This
initialization is thus needed to avoid reads of unitialized values.

New p3p algorithm (accepted by CVPR 2017) (#8301)

* add p3p source code

* indent 4

* update publication info

* fix filename

* interface done

* plug in done, test needed

* debugging

* for test

* a working version

* clean p3p code

* test

* test

* fix warning, blank line

* apply patch from @catree

* add reference info

* namespace, indent 4

* static solveQuartic

* put small functions to anonymous namespace

Same as https://github.com/opencv/opencv/pull/8525/commits/9210cefb36868e201f7e0c8b84e4ee4f91b5432b but for this file too.

 - check boundary strictly
 - initialize the variable before using it

- Fix StereoBM ROI (Region of interest) selection for stereo
images used by block matching algorithm

AffineBasedEstimator crashed when called with an existing CameraParameters.
This happens e.g. when using Stitcher in SCANS mode.
CameraraParameters is now cleared before any calculation is executed.

Use identity matrix if homography finding failed. Current behavior zeros out all points.

Update circlesgrid.cpp

Addressed comments

Update circlesgrid.cpp

removed whitespace

Fix #8177

In the previous version only the default stream was/could be used, i.e.
cv::cuda::Stream::Null().

With this change, HOG::compute() will now run in parallel over different
cuda::Streams.

The code has been reordered so that all data allocation is completed
first, then all the kernels are run in parallel over streams.

Fix #8177

Addressed comments.

This test case uses a matrix with more dimensions than columns. Without
the fix in
https://github.com/opencv/opencv/pull/8448/commits/b45e784bebe318826bb48a41023db3fe1660d91e
this crashes with a segmentation fault, hangs or simply fails with wrong
values.

Stitcher will now make a working copy of the CameraParams object to avoid side effects when composing Panorama.
Makes it possible to estimate transform once and then compose multiple panoramas. Useful for setup with fixed cameras.

This issue concerns only matrices with more dimensions than columns.
See https://github.com/opencv/opencv/issues/8447

- use suffixes like '.avx.cpp'
- added CMake-generated files for '.simd.hpp' optimization approach
- wrap HAL intrinsic headers into separate namespaces for different build flags
- automatic vzeroupper insertion (via CV_INSTRUMENT_REGION macro)

* export SVM::trainAuto to python #7224

* workaround for ABI compatibility of SVM::trainAuto

* add parameter comments to new SVM::trainAuto function

* Export ParamGrid member variables