This also fixes few clang build errors

some improvements of existing sse3 optimization of bilateral filter in case of 8uc3. Now perf tests take 6120ms instead of previous 7250ms (1.18x speed-up)

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

added recommended number of stripes to parallel_for_, modified some of the functions to use larger stripes (for better performance)

added GPU non local means brute force filter + tests

enabled SSE3 by default; integrated SSE3-optimized bilateral filter (by Grigoriy Frolov); modified API of non-local means (use Input/OutputArrays)

added special function to patch NaN's in image. call this function from bilateralfilter (should fix bug #1811)

Added threaded versions (enabled only for Tegra at the moment) of threshold, erode/dilate. Slighlty restructured the code of medianBlur.

replaced "const InputArray&" => "InputArray"; made InputArray and OutputArray references. added "None()" constant (no array()).

a big patch; use special proxy types (Input/OutputArray, Input/OutputArrayOfArrays) for passing in vectors, matrices etc.

added hconcat & vconcat functions for joining matrices; moved some inline functions out of the headers; fixed several bugs in documentation; removed MatND from docs