potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-29 23:33:05 +04:00
Code Packages Releases Wiki Activity

Merge pull request #24271 from Kumataro:fix24163

Browse Source 
Fix to convert float32 to int32/uint32 with rounding to nearest (ties to even). #24271

Fix https://github.com/opencv/opencv/issues/24163

### Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

- [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
- [x] There is a reference to the original bug report and related work
- [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
      Patch to opencv_extra has the same branch name.
- [x] The feature is well documented and sample code can be built with the project CMake

(carotene is BSD)
This commit is contained in:
Kumataro
2023-12-25 18:17:17 +09:00
committed by GitHub
parent d9d402916a
commit dba7186378
13 changed files with 323 additions and 164 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/core/include/opencv2/core/hal/intrin_neon.hpp
+3 -5
View File
@@ -1990,11 +1990,9 @@ inline v_int32x4 v_round(const v_float32x4& a)
#else
inline v_int32x4 v_round(const v_float32x4& a)
{
static const int32x4_t v_sign = vdupq_n_s32(1 << 31),
v_05 = vreinterpretq_s32_f32(vdupq_n_f32(0.5f));
int32x4_t v_addition = vorrq_s32(v_05, vandq_s32(v_sign, vreinterpretq_s32_f32(a.val)));
return v_int32x4(vcvtq_s32_f32(vaddq_f32(a.val, vreinterpretq_f32_s32(v_addition))));
// See https://github.com/opencv/opencv/pull/24271#issuecomment-1867318007
float32x4_t delta = vdupq_n_f32(12582912.0f);
return v_int32x4(vcvtq_s32_f32(vsubq_f32(vaddq_f32(a.val, delta), delta)));
}
#endif
inline v_int32x4 v_floor(const v_float32x4& a)