Merge pull request opencv#24205 from PeterJohnson:fix-msvc-arm64

ht_dec.c: Improve MSVC arm64 popcount performance opencv#24205

Use NEON instructions for ARM64 (implementation based on microsoft/STL#2127, which is Apache licensed).

Godbolt output here: https://godbolt.org/z/q7GPTqT14
Related patch to openjpeg: uclouvain/openjpeg#1479

### Pull Request Readiness Checklist

- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch

3rdparty/openjpeg/openjp2/ht_dec.c

@@ -55,6 +55,10 @@
 #define OPJ_COMPILER_GNUC
 #endif
 
+#if defined(OPJ_COMPILER_MSVC) && defined(_M_ARM64)
+#include <arm64_neon.h>
+#endif
+
 //************************************************************************/
 /** @brief Displays the error message for disabling the decoding of SPP and
   * MRP passes
@@ -71,6 +75,9 @@ OPJ_UINT32 population_count(OPJ_UINT32 val)
 {
 #if defined(OPJ_COMPILER_MSVC) && (defined(_M_IX86) || defined(_M_AMD64))
     return (OPJ_UINT32)__popcnt(val);
+#elif defined(OPJ_COMPILER_MSVC) && defined(_M_ARM64)
+    const __n64 temp = neon_cnt(__uint64ToN64_v(val));
+    return neon_addv8(temp).n8_i8[0];
 #elif (defined OPJ_COMPILER_GNUC)
     return (OPJ_UINT32)__builtin_popcount(val);
 #else