imgproc: dispatch color*

Lab/XYZ modes have been postponed (color_lab.cpp):
- need to split code for tables initialization and for pixels processing first
- no significant performance improvements for switching between SSE42 / AVX2 code generation

modules/core/include/opencv2/core/cv_cpu_dispatch.h

@@ -124,6 +124,10 @@
 
 #if defined CV_CPU_COMPILE_AVX && !defined CV_CPU_BASELINE_COMPILE_AVX
 struct VZeroUpperGuard {
+#ifdef __GNUC__
+    __attribute__((always_inline))
+#endif
+    inline VZeroUpperGuard() { _mm256_zeroupper(); }
 #ifdef __GNUC__
     __attribute__((always_inline))
 #endif

modules/core/include/opencv2/core/private.hpp

@@ -796,9 +796,9 @@ CV_EXPORTS InstrNode*   getCurrentNode();
 #endif
 
 #ifdef __CV_AVX_GUARD
-#define CV_INSTRUMENT_REGION(); __CV_AVX_GUARD CV_INSTRUMENT_REGION_();
+#define CV_INSTRUMENT_REGION() __CV_AVX_GUARD CV_INSTRUMENT_REGION_();
 #else
-#define CV_INSTRUMENT_REGION(); CV_INSTRUMENT_REGION_();
+#define CV_INSTRUMENT_REGION() CV_INSTRUMENT_REGION_();
 #endif
 
 namespace cv {