mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2024-11-25 11:40:44 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

corners

Browse Source
This commit is contained in:
Ilya Lavrenov 2014-09-24 22:33:23 +04:00
parent d30ce2b9ac
commit 12001a42f9
2 changed files with 23 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
modules/imgproc/src/canny.cpp
View File

@ -362,7 +362,7 @@ void cv::Canny( InputArray _src, OutputArray _dst,
}
}
#elif CV_NEON
for ( ; j < width - 8; j += 8)
for ( ; j <= width - 8; j += 8)
{
int16x8_t v_dx = vld1q_s16(_dx + j), v_dy = vld1q_s16(_dy + j);
vst1q_s32(_norm + j, vaddq_s32(vmovl_s16(vget_low_s16(v_dx)), vmovl_s16(vget_low_s16(v_dy))));
@ -394,7 +394,7 @@ void cv::Canny( InputArray _src, OutputArray _dst,
}
}
#elif CV_NEON
for ( ; j < width - 8; j += 8)
for ( ; j <= width - 8; j += 8)
{
int16x8_t v_dx = vld1q_s16(_dx + j), v_dy = vld1q_s16(_dy + j);
int32x4_t v_dxp = vmovl_s16(vget_low_s16(v_dx)), v_dyp = vmovl_s16(vget_low_s16(v_dy));

21
modules/imgproc/src/corner.cpp
View File

@ -146,6 +146,18 @@ static void calcHarris( const Mat& _cov, Mat& _dst, double k )
_mm_storeu_ps(dst + j, a);
}
}
#elif CV_NEON
float32x4_t v_k = vdupq_n_f32((float)k));
for( ; j <= size.width - 4; j += 4 )
{
float32x4x3_t v_src = vld3q_f32(cov + j + 3);
float32x4_t v_a = v_src.val[0], v_b = v_src.val[1], v_c = v_src.val[2];
float32x4_t v_ac_bb = vsubq_f32(vmulq_f32(v_a, v_c), vmulq_f32(v_b, v_b));
float32x4_t v_ac = vaddq_f32(v_a, v_c);
float32x4_t v_prod = vmulq_f32(v_k, vmulq_f32(v_ac, v_ac));
vst1q_f32(dst + j, vsubq_f32(v_ac_bb, v_prod));
}
#endif
for( ; j < size.width; j++ )
@ -641,6 +653,15 @@ void cv::preCornerDetect( InputArray _src, OutputArray _dst, int ksize, int bord
_mm_storeu_ps(dstdata + j, v_s1);
}
}
#elif CV_NEON
for( ; j <= size.width - 4; j += 4 )
{
float32x4_t v_dx = vld1q_f32(dxdata + j), v_dy = vld1q_f32(dydata + j);
float32x4_t v_s1 = vmulq_f32(v_dx, vmulq_f32(v_dx, vld1q_f32(d2ydata + j)));
float32x4_t v_s2 = vmulq_f32(v_dy, vmulq_f32(v_dy, vld1q_f32(d2xdata + j)));
float32x4_t v_s3 = vmulq_f32(v_dx, vmulq_f32(v_dy, vld1q_f32(dxydata + j)));
vst1q_f32(dstdata + j, vaddq_f32(vaddq_f32(v_s1, v_s2), vmulq_n_f32(v_s3, -2.0f)));
}
#endif
for( ; j < size.width; j++ )