Skip to content

O
opencv
Commit 27cddfb8 authored by Alexander Alekhin's avatar Alexander Alekhin
Browse files
Options

Merge pull request #10661 from alalek:parallel_kmeans

parents 3f116468 90aac764
Hide whitespace changes
Inline Side-by-side
Showing with 256 additions and 195 deletions
modules/core/perf/perf_math.cpp
View file @ 27cddfb8
......@@ -3,13 +3,10 @@
using namespace std;
using namespace cv;
using namespace perf;
using std::tr1::make_tuple;
using std::tr1::get;
typedef perf::TestBaseWithParam<size_t> VectorLength;
namespace {
typedef std::tr1::tuple<int, int> MaxDim_MaxPoints_t;
typedef perf::TestBaseWithParam<MaxDim_MaxPoints_t> MaxDim_MaxPoints;
typedef perf::TestBaseWithParam<size_t> VectorLength;
PERF_TEST_P(VectorLength, phase32f, testing::Values(128, 1000, 128*1024, 512*1024, 1024*1024))
{
......@@ -39,47 +36,125 @@ PERF_TEST_P(VectorLength, phase64f, testing::Values(128, 1000, 128*1024, 512*102
SANITY_CHECK(angle, 5e-5);
}
PERF_TEST_P( MaxDim_MaxPoints, kmeans,
testing::Combine( testing::Values( 16, 32, 64 ),
testing::Values( 300, 400, 500) ) )
typedef perf::TestBaseWithParam< testing::tuple<int, int, int> > KMeans;
PERF_TEST_P_(KMeans, single_iter)
{
RNG& rng = theRNG();
const int MAX_DIM = get<0>(GetParam());
const int MAX_POINTS = get<1>(GetParam());
const int K = testing::get<0>(GetParam());
const int dims = testing::get<1>(GetParam());
const int N = testing::get<2>(GetParam());
const int attempts = 5;
Mat labels, centers;
int i, N = 0, N0 = 0, K = 0, dims = 0;
dims = rng.uniform(1, MAX_DIM+1);
N = rng.uniform(1, MAX_POINTS+1);
N0 = rng.uniform(1, MAX(N/10, 2));
K = rng.uniform(1, N+1);
Mat data(N, dims, CV_32F);
rng.fill(data, RNG::UNIFORM, -0.1, 0.1);
const int N0 = K;
Mat data0(N0, dims, CV_32F);
rng.fill(data0, RNG::UNIFORM, -1, 1);
for (int i = 0; i < N; i++)
{
int base = rng.uniform(0, N0);
cv::add(data0.row(base), data.row(i), data.row(i));
}
declare.in(data);
Mat labels, centers;
TEST_CYCLE()
{
kmeans(data, K, labels, TermCriteria(TermCriteria::MAX_ITER+TermCriteria::EPS, 1, 0),
attempts, KMEANS_PP_CENTERS, centers);
}
SANITY_CHECK_NOTHING();
}
PERF_TEST_P_(KMeans, good)
{
RNG& rng = theRNG();
const int K = testing::get<0>(GetParam());
const int dims = testing::get<1>(GetParam());
const int N = testing::get<2>(GetParam());
const int attempts = 5;
Mat data(N, dims, CV_32F);
for( i = 0; i < N; i++ )
data0.row(rng.uniform(0, N0)).copyTo(data.row(i));
rng.fill(data, RNG::UNIFORM, -0.1, 0.1);
const int N0 = K;
Mat data0(N0, dims, CV_32F);
rng.fill(data0, RNG::UNIFORM, -1, 1);
for (int i = 0; i < N; i++)
{
int base = rng.uniform(0, N0);
cv::add(data0.row(base), data.row(i), data.row(i));
}
declare.in(data);
Mat labels, centers;
TEST_CYCLE()
{
kmeans(data, K, labels, TermCriteria(TermCriteria::MAX_ITER+TermCriteria::EPS, 30, 0),
attempts, KMEANS_PP_CENTERS, centers);
}
Mat clusterPointsNumber = Mat::zeros(1, K, CV_32S);
SANITY_CHECK_NOTHING();
}
PERF_TEST_P_(KMeans, with_duplicates)
{
RNG& rng = theRNG();
const int K = testing::get<0>(GetParam());
const int dims = testing::get<1>(GetParam());
const int N = testing::get<2>(GetParam());
const int attempts = 5;
Mat data(N, dims, CV_32F, Scalar::all(0));
const int N0 = std::max(2, K * 2 / 3);
Mat data0(N0, dims, CV_32F);
rng.fill(data0, RNG::UNIFORM, -1, 1);
for (int i = 0; i < N; i++)
{
int base = rng.uniform(0, N0);
data0.row(base).copyTo(data.row(i));
}
declare.in(data);
for( i = 0; i < labels.rows; i++ )
Mat labels, centers;
TEST_CYCLE()
{
int clusterIdx = labels.at<int>(i);
clusterPointsNumber.at<int>(clusterIdx)++;
kmeans(data, K, labels, TermCriteria(TermCriteria::MAX_ITER+TermCriteria::EPS, 30, 0),
attempts, KMEANS_PP_CENTERS, centers);
}
Mat sortedClusterPointsNumber;
cv::sort(clusterPointsNumber, sortedClusterPointsNumber, cv::SORT_EVERY_ROW + cv::SORT_ASCENDING);
SANITY_CHECK_NOTHING();
}
INSTANTIATE_TEST_CASE_P(/*nothing*/ , KMeans,
testing::Values(
// K clusters, dims, N points
testing::make_tuple(2, 3, 100000),
testing::make_tuple(4, 3, 500),
testing::make_tuple(4, 3, 1000),
testing::make_tuple(4, 3, 10000),
testing::make_tuple(8, 3, 1000),
testing::make_tuple(8, 16, 1000),
testing::make_tuple(8, 64, 1000),
testing::make_tuple(16, 16, 1000),
testing::make_tuple(16, 32, 1000),
testing::make_tuple(32, 16, 1000),
testing::make_tuple(32, 32, 1000),
testing::make_tuple(100, 2, 1000)
)
);
SANITY_CHECK(sortedClusterPointsNumber);
}
modules/core/src/kmeans.cpp
View file @ 27cddfb8
......@@ -10,7 +10,7 @@
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2000-2018, Intel Corporation, all rights reserved.
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
// Copyright (C) 2013, OpenCV Foundation, all rights reserved.
// Third party copyrights are property of their respective owners.
......@@ -42,105 +42,100 @@
//M*/
#include "precomp.hpp"
#include <opencv2/core/utils/configuration.private.hpp>
////////////////////////////////////////// kmeans ////////////////////////////////////////////
namespace cv
{
static void generateRandomCenter(const std::vector<Vec2f>& box, float* center, RNG& rng)
static int CV_KMEANS_PARALLEL_GRANULARITY = (int)utils::getConfigurationParameterSizeT("OPENCV_KMEANS_PARALLEL_GRANULARITY", 1000);
static void generateRandomCenter(int dims, const Vec2f* box, float* center, RNG& rng)
{
size_t j, dims = box.size();
float margin = 1.f/dims;
for( j = 0; j < dims; j++ )
for (int j = 0; j < dims; j++)
center[j] = ((float)rng*(1.f+margin*2.f)-margin)*(box[j][1] - box[j][0]) + box[j][0];
}
class KMeansPPDistanceComputer : public ParallelLoopBody
{
public:
KMeansPPDistanceComputer( float *_tdist2,
const float *_data,
const float *_dist,
int _dims,
size_t _step,
size_t _stepci )
: tdist2(_tdist2),
data(_data),
dist(_dist),
dims(_dims),
step(_step),
stepci(_stepci) { }
KMeansPPDistanceComputer(float *tdist2_, const Mat& data_, const float *dist_, int ci_) :
tdist2(tdist2_), data(data_), dist(dist_), ci(ci_)
{ }
void operator()( const cv::Range& range ) const
{
CV_TRACE_FUNCTION();
const int begin = range.start;
const int end = range.end;
const int dims = data.cols;
for ( int i = begin; i<end; i++ )
for (int i = begin; i<end; i++)
{
tdist2[i] = std::min(normL2Sqr(data + step*i, data + stepci, dims), dist[i]);
tdist2[i] = std::min(normL2Sqr(data.ptr<float>(i), data.ptr<float>(ci), dims), dist[i]);
}
}
private:
KMeansPPDistanceComputer& operator=(const KMeansPPDistanceComputer&); // to quiet MSVC
KMeansPPDistanceComputer& operator=(const KMeansPPDistanceComputer&); // = delete
float *tdist2;
const float *data;
const Mat& data;
const float *dist;
const int dims;
const size_t step;
const size_t stepci;
const int ci;
};
/*
k-means center initialization using the following algorithm:
Arthur & Vassilvitskii (2007) k-means++: The Advantages of Careful Seeding
*/
static void generateCentersPP(const Mat& _data, Mat& _out_centers,
static void generateCentersPP(const Mat& data, Mat& _out_centers,
int K, RNG& rng, int trials)
{
CV_TRACE_FUNCTION();
int i, j, k, dims = _data.cols, N = _data.rows;
const float* data = _data.ptr<float>(0);
size_t step = _data.step/sizeof(data[0]);
std::vector<int> _centers(K);
const int dims = data.cols, N = data.rows;
cv::AutoBuffer<int, 64> _centers(K);
int* centers = &_centers[0];
std::vector<float> _dist(N*3);
cv::AutoBuffer<float, 0> _dist(N*3);
float* dist = &_dist[0], *tdist = dist + N, *tdist2 = tdist + N;
double sum0 = 0;
centers[0] = (unsigned)rng % N;
for( i = 0; i < N; i++ )
for (int i = 0; i < N; i++)
{
dist[i] = normL2Sqr(data + step*i, data + step*centers[0], dims);
dist[i] = normL2Sqr(data.ptr<float>(i), data.ptr<float>(centers[0]), dims);
sum0 += dist[i];
}
for( k = 1; k < K; k++ )
for (int k = 1; k < K; k++)
{
double bestSum = DBL_MAX;
int bestCenter = -1;
for( j = 0; j < trials; j++ )
for (int j = 0; j < trials; j++)
{
double p = (double)rng*sum0, s = 0;
for( i = 0; i < N-1; i++ )
if( (p -= dist[i]) <= 0 )
double p = (double)rng*sum0;
int ci = 0;
for (; ci < N - 1; ci++)
{
p -= dist[ci];
if (p <= 0)
break;
int ci = i;
}
parallel_for_(Range(0, N),
KMeansPPDistanceComputer(tdist2, data, dist, dims, step, step*ci));
for( i = 0; i < N; i++ )
KMeansPPDistanceComputer(tdist2, data, dist, ci),
divUp(dims * N, CV_KMEANS_PARALLEL_GRANULARITY));
double s = 0;
for (int i = 0; i < N; i++)
{
s += tdist2[i];
}
if( s < bestSum )
if (s < bestSum)
{
bestSum = s;
bestCenter = ci;
......@@ -152,39 +147,39 @@ static void generateCentersPP(const Mat& _data, Mat& _out_centers,
std::swap(dist, tdist);
}
for( k = 0; k < K; k++ )
for (int k = 0; k < K; k++)
{
const float* src = data + step*centers[k];
const float* src = data.ptr<float>(centers[k]);
float* dst = _out_centers.ptr<float>(k);
for( j = 0; j < dims; j++ )
for (int j = 0; j < dims; j++)
dst[j] = src[j];
}
}
template<bool onlyDistance>
class KMeansDistanceComputer : public ParallelLoopBody
{
public:
KMeansDistanceComputer( double *_distances,
int *_labels,
const Mat& _data,
const Mat& _centers,
bool _onlyDistance = false )
: distances(_distances),
labels(_labels),
data(_data),
centers(_centers),
onlyDistance(_onlyDistance)
KMeansDistanceComputer( double *distances_,
int *labels_,
const Mat& data_,
const Mat& centers_)
: distances(distances_),
labels(labels_),
data(data_),
centers(centers_)
{
}
void operator()( const Range& range ) const
{
CV_TRACE_FUNCTION();
const int begin = range.start;
const int end = range.end;
const int K = centers.rows;
const int dims = centers.cols;
for( int i = begin; i<end; ++i)
for (int i = begin; i < end; ++i)
{
const float *sample = data.ptr<float>(i);
if (onlyDistance)
......@@ -193,34 +188,36 @@ public:
distances[i] = normL2Sqr(sample, center, dims);
continue;
}
int k_best = 0;
double min_dist = DBL_MAX;
for( int k = 0; k < K; k++ )
else
{
const float* center = centers.ptr<float>(k);
const double dist = normL2Sqr(sample, center, dims);
int k_best = 0;
double min_dist = DBL_MAX;
if( min_dist > dist )
for (int k = 0; k < K; k++)
{
min_dist = dist;
k_best = k;
const float* center = centers.ptr<float>(k);
const double dist = normL2Sqr(sample, center, dims);
if (min_dist > dist)
{
min_dist = dist;
k_best = k;
}
}
}
distances[i] = min_dist;
labels[i] = k_best;
distances[i] = min_dist;
labels[i] = k_best;
}
}
}
private:
KMeansDistanceComputer& operator=(const KMeansDistanceComputer&); // to quiet MSVC
KMeansDistanceComputer& operator=(const KMeansDistanceComputer&); // = delete
double *distances;
int *labels;
const Mat& data;
const Mat& centers;
bool onlyDistance;
};
}
......@@ -231,13 +228,12 @@ double cv::kmeans( InputArray _data, int K,
int flags, OutputArray _centers )
{
CV_INSTRUMENT_REGION()
const int SPP_TRIALS = 3;
Mat data0 = _data.getMat();
bool isrow = data0.rows == 1;
int N = isrow ? data0.cols : data0.rows;
int dims = (isrow ? 1 : data0.cols)*data0.channels();
int type = data0.depth();
const bool isrow = data0.rows == 1;
const int N = isrow ? data0.cols : data0.rows;
const int dims = (isrow ? 1 : data0.cols)*data0.channels();
const int type = data0.depth();
attempts = std::max(attempts, 1);
CV_Assert( data0.dims <= 2 && type == CV_32F && K > 0 );
......@@ -248,129 +244,115 @@ double cv::kmeans( InputArray _data, int K,
_bestLabels.create(N, 1, CV_32S, -1, true);
Mat _labels, best_labels = _bestLabels.getMat();
if( flags & CV_KMEANS_USE_INITIAL_LABELS )
if (flags & CV_KMEANS_USE_INITIAL_LABELS)
{
CV_Assert( (best_labels.cols == 1 || best_labels.rows == 1) &&
best_labels.cols*best_labels.rows == N &&
best_labels.type() == CV_32S &&
best_labels.isContinuous());
best_labels.copyTo(_labels);
best_labels.reshape(1, N).copyTo(_labels);
for (int i = 0; i < N; i++)
{
CV_Assert((unsigned)_labels.at<int>(i) < (unsigned)K);
}
}
else
{
if( !((best_labels.cols == 1 || best_labels.rows == 1) &&
if (!((best_labels.cols == 1 || best_labels.rows == 1) &&
best_labels.cols*best_labels.rows == N &&
best_labels.type() == CV_32S &&
best_labels.isContinuous()))
best_labels.create(N, 1, CV_32S);
best_labels.type() == CV_32S &&
best_labels.isContinuous()))
{
_bestLabels.create(N, 1, CV_32S);
best_labels = _bestLabels.getMat();
}
_labels.create(best_labels.size(), best_labels.type());
}
int* labels = _labels.ptr<int>();
Mat centers(K, dims, type), old_centers(K, dims, type), temp(1, dims, type);
std::vector<int> counters(K);
std::vector<Vec2f> _box(dims);
Mat dists(1, N, CV_64F);
Vec2f* box = &_box[0];
double best_compactness = DBL_MAX, compactness = 0;
cv::AutoBuffer<int, 64> counters(K);
cv::AutoBuffer<double, 64> dists(N);
RNG& rng = theRNG();
int a, iter, i, j, k;
if( criteria.type & TermCriteria::EPS )
if (criteria.type & TermCriteria::EPS)
criteria.epsilon = std::max(criteria.epsilon, 0.);
else
criteria.epsilon = FLT_EPSILON;
criteria.epsilon *= criteria.epsilon;
if( criteria.type & TermCriteria::COUNT )
if (criteria.type & TermCriteria::COUNT)
criteria.maxCount = std::min(std::max(criteria.maxCount, 2), 100);
else
criteria.maxCount = 100;
if( K == 1 )
if (K == 1)
{
attempts = 1;
criteria.maxCount = 2;
}
const float* sample = data.ptr<float>(0);
for( j = 0; j < dims; j++ )
box[j] = Vec2f(sample[j], sample[j]);
for( i = 1; i < N; i++ )
cv::AutoBuffer<Vec2f, 64> box(dims);
if (!(flags & KMEANS_PP_CENTERS))
{
sample = data.ptr<float>(i);
for( j = 0; j < dims; j++ )
{
float v = sample[j];
box[j][0] = std::min(box[j][0], v);
box[j][1] = std::max(box[j][1], v);
const float* sample = data.ptr<float>(0);
for (int j = 0; j < dims; j++)
box[j] = Vec2f(sample[j], sample[j]);
}
for (int i = 1; i < N; i++)
{
const float* sample = data.ptr<float>(i);
for (int j = 0; j < dims; j++)
{
float v = sample[j];
box[j][0] = std::min(box[j][0], v);
box[j][1] = std::max(box[j][1], v);
}
}
}
for( a = 0; a < attempts; a++ )
double best_compactness = DBL_MAX;
for (int a = 0; a < attempts; a++)
{
double max_center_shift = DBL_MAX;
for( iter = 0;; )
double compactness = 0;
for (int iter = 0; ;)
{
double max_center_shift = iter == 0 ? DBL_MAX : 0.0;
swap(centers, old_centers);
if( iter == 0 && (a > 0 || !(flags & KMEANS_USE_INITIAL_LABELS)) )
if (iter == 0 && (a > 0 || !(flags & KMEANS_USE_INITIAL_LABELS)))
{
if( flags & KMEANS_PP_CENTERS )
if (flags & KMEANS_PP_CENTERS)
generateCentersPP(data, centers, K, rng, SPP_TRIALS);
else
{
for( k = 0; k < K; k++ )
generateRandomCenter(_box, centers.ptr<float>(k), rng);
for (int k = 0; k < K; k++)
generateRandomCenter(dims, box, centers.ptr<float>(k), rng);
}
}
else
{
if( iter == 0 && a == 0 && (flags & KMEANS_USE_INITIAL_LABELS) )
{
for( i = 0; i < N; i++ )
CV_Assert( (unsigned)labels[i] < (unsigned)K );
}
// compute centers
centers = Scalar(0);
for( k = 0; k < K; k++ )
for (int k = 0; k < K; k++)
counters[k] = 0;
for( i = 0; i < N; i++ )
for (int i = 0; i < N; i++)
{
sample = data.ptr<float>(i);
k = labels[i];
const float* sample = data.ptr<float>(i);
int k = labels[i];
float* center = centers.ptr<float>(k);
j=0;
#if CV_ENABLE_UNROLLED
for(; j <= dims - 4; j += 4 )
{
float t0 = center[j] + sample[j];
float t1 = center[j+1] + sample[j+1];
center[j] = t0;
center[j+1] = t1;
t0 = center[j+2] + sample[j+2];
t1 = center[j+3] + sample[j+3];
center[j+2] = t0;
center[j+3] = t1;
}
#endif
for( ; j < dims; j++ )
for (int j = 0; j < dims; j++)
center[j] += sample[j];
counters[k]++;
}
if( iter > 0 )
max_center_shift = 0;
for( k = 0; k < K; k++ )
for (int k = 0; k < K; k++)
{
if( counters[k] != 0 )
if (counters[k] != 0)
continue;
// if some cluster appeared to be empty then:
......@@ -378,29 +360,28 @@ double cv::kmeans( InputArray _data, int K,
// 2. find the farthest from the center point in the biggest cluster
// 3. exclude the farthest point from the biggest cluster and form a new 1-point cluster.
int max_k = 0;
for( int k1 = 1; k1 < K; k1++ )
for (int k1 = 1; k1 < K; k1++)
{
if( counters[max_k] < counters[k1] )
if (counters[max_k] < counters[k1])
max_k = k1;
}
double max_dist = 0;
int farthest_i = -1;
float* new_center = centers.ptr<float>(k);
float* old_center = centers.ptr<float>(max_k);
float* _old_center = temp.ptr<float>(); // normalized
float* base_center = centers.ptr<float>(max_k);
float* _base_center = temp.ptr<float>(); // normalized
float scale = 1.f/counters[max_k];
for( j = 0; j < dims; j++ )
_old_center[j] = old_center[j]*scale;
for (int j = 0; j < dims; j++)
_base_center[j] = base_center[j]*scale;
for( i = 0; i < N; i++ )
for (int i = 0; i < N; i++)
{
if( labels[i] != max_k )
if (labels[i] != max_k)
continue;
sample = data.ptr<float>(i);
double dist = normL2Sqr(sample, _old_center, dims);
const float* sample = data.ptr<float>(i);
double dist = normL2Sqr(sample, _base_center, dims);
if( max_dist <= dist )
if (max_dist <= dist)
{
max_dist = dist;
farthest_i = i;
......@@ -410,29 +391,30 @@ double cv::kmeans( InputArray _data, int K,
counters[max_k]--;
counters[k]++;
labels[farthest_i] = k;
sample = data.ptr<float>(farthest_i);
for( j = 0; j < dims; j++ )
const float* sample = data.ptr<float>(farthest_i);
float* cur_center = centers.ptr<float>(k);
for (int j = 0; j < dims; j++)
{
old_center[j] -= sample[j];
new_center[j] += sample[j];
base_center[j] -= sample[j];
cur_center[j] += sample[j];
}
}
for( k = 0; k < K; k++ )
for (int k = 0; k < K; k++)
{
float* center = centers.ptr<float>(k);
CV_Assert( counters[k] != 0 );
float scale = 1.f/counters[k];
for( j = 0; j < dims; j++ )
for (int j = 0; j < dims; j++)
center[j] *= scale;
if( iter > 0 )
if (iter > 0)
{
double dist = 0;
const float* old_center = old_centers.ptr<float>(k);
for( j = 0; j < dims; j++ )
for (int j = 0; j < dims; j++)
{
double t = center[j] - old_center[j];
dist += t*t;
......@@ -444,25 +426,29 @@ double cv::kmeans( InputArray _data, int K,
bool isLastIter = (++iter == MAX(criteria.maxCount, 2) || max_center_shift <= criteria.epsilon);
// assign labels
dists = 0;
double* dist = dists.ptr<double>(0);
parallel_for_(Range(0, N), KMeansDistanceComputer(dist, labels, data, centers, isLastIter));
compactness = sum(dists)[0];
if (isLastIter)
{
// don't re-assign labels to avoid creation of empty clusters
parallel_for_(Range(0, N), KMeansDistanceComputer<true>(dists, labels, data, centers), divUp(dims * N, CV_KMEANS_PARALLEL_GRANULARITY));
compactness = sum(Mat(Size(N, 1), CV_64F, &dists[0]))[0];
break;
}
else
{
// assign labels
parallel_for_(Range(0, N), KMeansDistanceComputer<false>(dists, labels, data, centers), divUp(dims * N * K, CV_KMEANS_PARALLEL_GRANULARITY));
}
}
if( compactness < best_compactness )
if (compactness < best_compactness)
{
best_compactness = compactness;
if( _centers.needed() )
if (_centers.needed())
{
Mat reshaped = centers;
if(_centers.fixedType() && _centers.channels() == dims)
reshaped = centers.reshape(dims);
reshaped.copyTo(_centers);
if (_centers.fixedType() && _centers.channels() == dims)
centers.reshape(dims).copyTo(_centers);
else
centers.copyTo(_centers);
}
_labels.copyTo(best_labels);
}
......
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