Skip to content

O
opencv
Commit 77dda061 authored by Vadim Pisarevsky's avatar Vadim Pisarevsky
Browse files
Options

fixed fastAtan2 and cardToPolar accuracy (thanks to Andrey Kamaev)

parent 8989e0b0
Hide whitespace changes
Inline Side-by-side
Showing with 79 additions and 53 deletions
modules/core/src/mathfuncs.cpp
View file @ 77dda061
......@@ -49,22 +49,38 @@ namespace cv
static const int MAX_BLOCK_SIZE = 1024;
typedef void (*MathFunc)(const void* src, void* dst, int len);
static const float atan2_p1 = 0.9997878412794807f*(float)(180/CV_PI);
static const float atan2_p3 = -0.3258083974640975f*(float)(180/CV_PI);
static const float atan2_p5 = 0.1555786518463281f*(float)(180/CV_PI);
static const float atan2_p7 = -0.04432655554792128f*(float)(180/CV_PI);
float fastAtan2( float y, float x )
{
double a, x2 = (double)x*x, y2 = (double)y*y;
if( y2 <= x2 )
{
a = (180./CV_PI)*x*y*(x2 + 0.43157974*y2)/(x2*x2 + y2*(0.76443945*x2 + 0.05831938*y2) + DBL_EPSILON);
return (float)(x < 0 ? a + 180 : y >= 0 ? a : 360+a);
}
a = (180./CV_PI)*x*y*(y2 + 0.43157974*x2)/(y2*y2 + x2*(0.76443945*y2 + 0.05831938*x2) + DBL_EPSILON);
return (float)(y >= 0 ? 90 - a : 270 - a);
float ax = std::abs(x), ay = std::abs(y);
float a, c, c2;
if( ax >= ay )
{
c = ay/(ax + (float)DBL_EPSILON);
c2 = c*c;
a = (((atan2_p7*c2 + atan2_p5)*c2 + atan2_p3)*c2 + atan2_p1)*c;
}
else
{
c = ax/(ay + (float)DBL_EPSILON);
c2 = c*c;
a = 90.f - (((atan2_p7*c2 + atan2_p5)*c2 + atan2_p3)*c2 + atan2_p1)*c;
}
if( x < 0 )
a = 180.f - a;
if( y < 0 )
a = 360.f - a;
return a;
}
static void FastAtan2_32f(const float *Y, const float *X, float *angle, int len, bool angleInDegrees=true )
{
int i = 0;
float scale = angleInDegrees ? (float)(180/CV_PI) : 1.f;
float scale = angleInDegrees ? 1 : (float)(CV_PI/180);
#ifdef HAVE_TEGRA_OPTIMIZATION
if (tegra::FastAtan2_32f(Y, X, angle, len, scale))
......@@ -72,54 +88,66 @@ static void FastAtan2_32f(const float *Y, const float *X, float *angle, int len,
#endif
#if CV_SSE2
if( USE_SSE2 )
{
Cv32suf iabsmask; iabsmask.i = 0x7fffffff;
__m128 eps = _mm_set1_ps((float)DBL_EPSILON), absmask = _mm_set1_ps(iabsmask.f);
__m128 _90 = _mm_set1_ps((float)(CV_PI*0.5)), _180 = _mm_set1_ps((float)CV_PI), _360 = _mm_set1_ps((float)(CV_PI*2));
__m128 zero = _mm_setzero_ps(), scale4 = _mm_set1_ps(scale);
__m128 p0 = _mm_set1_ps(0.43157974f), q0 = _mm_set1_ps(0.76443945f), q1 = _mm_set1_ps(0.05831938f);
if( USE_SSE2 )
{
Cv32suf iabsmask; iabsmask.i = 0x7fffffff;
__m128 eps = _mm_set1_ps((float)DBL_EPSILON), absmask = _mm_set1_ps(iabsmask.f);
__m128 _90 = _mm_set1_ps(90.f), _180 = _mm_set1_ps(180.f), _360 = _mm_set1_ps(360.f);
__m128 z = _mm_setzero_ps(), scale4 = _mm_set1_ps(scale);
__m128 p1 = _mm_set1_ps(atan2_p1), p3 = _mm_set1_ps(atan2_p3);
__m128 p5 = _mm_set1_ps(atan2_p5), p7 = _mm_set1_ps(atan2_p7);
for( ; i <= len - 4; i += 4 )
{
__m128 x4 = _mm_loadu_ps(X + i), y4 = _mm_loadu_ps(Y + i);
__m128 xq4 = _mm_mul_ps(x4, x4), yq4 = _mm_mul_ps(y4, y4);
__m128 xly = _mm_cmplt_ps(xq4, yq4);
__m128 t = _mm_min_ps(xq4, yq4);
xq4 = _mm_max_ps(xq4, yq4); yq4 = t;
__m128 z4 = _mm_div_ps(_mm_mul_ps(_mm_mul_ps(x4, y4), _mm_add_ps(xq4, _mm_mul_ps(yq4, p0))),
_mm_add_ps(eps, _mm_add_ps(_mm_mul_ps(xq4, xq4),
_mm_mul_ps(yq4, _mm_add_ps(_mm_mul_ps(xq4, q0),
_mm_mul_ps(yq4, q1))))));
// a4 <- x < y ? 90 : 0;
__m128 a4 = _mm_and_ps(xly, _90);
// a4 <- (y < 0 ? 360 - a4 : a4) == ((x < y ? y < 0 ? 270 : 90) : (y < 0 ? 360 : 0))
__m128 mask = _mm_cmplt_ps(y4, zero);
a4 = _mm_or_ps(_mm_and_ps(_mm_sub_ps(_360, a4), mask), _mm_andnot_ps(mask, a4));
// a4 <- (x < 0 && !(x < y) ? 180 : a4)
mask = _mm_andnot_ps(xly, _mm_cmplt_ps(x4, zero));
a4 = _mm_or_ps(_mm_and_ps(_180, mask), _mm_andnot_ps(mask, a4));
// a4 <- (x < y ? a4 - z4 : a4 + z4)
a4 = _mm_mul_ps(_mm_add_ps(_mm_xor_ps(z4, _mm_andnot_ps(absmask, xly)), a4), scale4);
_mm_storeu_ps(angle + i, a4);
}
for( ; i <= len - 4; i += 4 )
{
__m128 x = _mm_loadu_ps(X + i), y = _mm_loadu_ps(Y + i);
__m128 ax = _mm_and_ps(x, absmask), ay = _mm_and_ps(y, absmask);
__m128 mask = _mm_cmplt_ps(ax, ay);
__m128 tmin = _mm_min_ps(ax, ay), tmax = _mm_max_ps(ax, ay);
__m128 c = _mm_div_ps(tmin, _mm_add_ps(tmax, eps));
__m128 c2 = _mm_mul_ps(c, c);
__m128 a = _mm_mul_ps(c2, p7);
a = _mm_mul_ps(_mm_add_ps(a, p5), c2);
a = _mm_mul_ps(_mm_add_ps(a, p3), c2);
a = _mm_mul_ps(_mm_add_ps(a, p1), c);
__m128 b = _mm_sub_ps(_90, a);
a = _mm_xor_ps(a, _mm_and_ps(_mm_xor_ps(a, b), mask));
b = _mm_sub_ps(_180, a);
mask = _mm_cmplt_ps(x, z);
a = _mm_xor_ps(a, _mm_and_ps(_mm_xor_ps(a, b), mask));
b = _mm_sub_ps(_360, a);
mask = _mm_cmplt_ps(y, z);
a = _mm_xor_ps(a, _mm_and_ps(_mm_xor_ps(a, b), mask));
a = _mm_mul_ps(a, scale4);
_mm_storeu_ps(angle + i, a);
}
}
#endif
for( ; i < len; i++ )
{
double x = X[i], y = Y[i], x2 = x*x, y2 = y*y, a;
if( y2 <= x2 )
a = (x < 0 ? CV_PI : y >= 0 ? 0 : CV_PI*2) +
x*y*(x2 + 0.43157974*y2)/(x2*x2 + y2*(0.76443945*x2 + 0.05831938*y2) + (float)DBL_EPSILON);
float x = X[i], y = Y[i];
float ax = std::abs(x), ay = std::abs(y);
float a, c, c2;
if( ax >= ay )
{
c = ay/(ax + (float)DBL_EPSILON);
c2 = c*c;
a = (((atan2_p7*c2 + atan2_p5)*c2 + atan2_p3)*c2 + atan2_p1)*c;
}
else
{
a = (y >= 0 ? CV_PI*0.5 : CV_PI*1.5) -
x*y*(y2 + 0.43157974*x2)/(y2*y2 + x2*(0.76443945*y2 + 0.05831938*x2) + (float)DBL_EPSILON);
c = ax/(ay + (float)DBL_EPSILON);
c2 = c*c;
a = 90.f - (((atan2_p7*c2 + atan2_p5)*c2 + atan2_p3)*c2 + atan2_p1)*c;
}
if( x < 0 )
a = 180.f - a;
if( y < 0 )
a = 360.f - a;
angle[i] = (float)(a*scale);
}
}
......
modules/core/src/system.cpp
View file @ 77dda061
......@@ -193,20 +193,18 @@ bool checkHardwareSupport(int feature)
return currentFeatures->have[feature];
}
#ifdef HAVE_IPP
volatile bool useOptimizedFlag = true;
volatile bool useOptimizedFlag = true;
#ifdef HAVE_IPP
struct IPPInitializer
{
IPPInitializer(void) { ippStaticInit(); }
};
IPPInitializer ippInitializer;
#else
volatile bool useOptimizedFlag = true;
#endif
volatile bool USE_SSE2 = false;
volatile bool USE_SSE2 = featuresEnabled.have[CV_CPU_SSE2];
void setUseOptimized( bool flag )
{
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment