Skip to content

O
opencv
Commit f678c8f0 authored by Vadim Pisarevsky's avatar Vadim Pisarevsky
Browse files
Options

made FLANN (and dependent on it code) build fine on Windows. Most of the changes… 

made FLANN (and dependent on it code) build fine on Windows. Most of the changes are added CV_EXPORTS' into the class declarations and extern variables are turned into external functions (as a nice side effect the variables are now read-only)
parent d611fb61
Show whitespace changes
Inline Side-by-side
Showing with 159 additions and 147 deletions
modules/flann/include/opencv2/flann/allocator.h
View file @ f678c8f0
......@@ -70,7 +70,7 @@ T* allocate(size_t count = 1)
const size_t WORDSIZE=16;
const size_t BLOCKSIZE=8192;
class PooledAllocator
class CV_EXPORTS PooledAllocator
{
/* We maintain memory alignment to word boundaries by requiring that all
allocations be in multiples of the machine wordsize. */
......
modules/flann/include/opencv2/flann/autotuned_index.h
View file @ f678c8f0
......@@ -59,11 +59,11 @@ struct AutotunedIndexParams : public IndexParams {
void print() const
{
logger.info("Index type: %d\n",(int)algorithm);
logger.info("Target precision: %g\n", target_precision);
logger.info("Build weight: %g\n", build_weight);
logger.info("Memory weight: %g\n", memory_weight);
logger.info("Sample fraction: %g\n", sample_fraction);
logger().info("Index type: %d\n",(int)algorithm);
logger().info("logger(). precision: %g\n", target_precision);
logger().info("Build weight: %g\n", build_weight);
logger().info("Memory weight: %g\n", memory_weight);
logger().info("Sample fraction: %g\n", sample_fraction);
}
};
......@@ -117,10 +117,10 @@ public:
virtual void buildIndex()
{
bestParams = estimateBuildParams();
logger.info("----------------------------------------------------\n");
logger.info("Autotuned parameters:\n");
logger().info("----------------------------------------------------\n");
logger().info("Autotuned parameters:\n");
bestParams->print();
logger.info("----------------------------------------------------\n");
logger().info("----------------------------------------------------\n");
flann_algorithm_t index_type = bestParams->getIndexType();
switch (index_type) {
case LINEAR:
......@@ -234,7 +234,7 @@ private:
int checks;
const int nn = 1;
logger.info("KMeansTree using params: max_iterations=%d, branching=%d\n", kmeans_params.iterations, kmeans_params.branching);
logger().info("KMeansTree using params: max_iterations=%d, branching=%d\n", kmeans_params.iterations, kmeans_params.branching);
KMeansIndex<ELEM_TYPE> kmeans(sampledDataset, kmeans_params);
// measure index build time
t.start();
......@@ -250,7 +250,7 @@ private:
cost.searchTimeCost = searchTime;
cost.buildTimeCost = buildTime;
cost.timeCost = (buildTime*index_params.build_weight+searchTime);
logger.info("KMeansTree buildTime=%g, searchTime=%g, timeCost=%g, buildTimeFactor=%g\n",buildTime, searchTime, cost.timeCost, index_params.build_weight);
logger().info("KMeansTree buildTime=%g, searchTime=%g, timeCost=%g, buildTimeFactor=%g\n",buildTime, searchTime, cost.timeCost, index_params.build_weight);
}
......@@ -260,7 +260,7 @@ private:
int checks;
const int nn = 1;
logger.info("KDTree using params: trees=%d\n",kdtree_params.trees);
logger().info("KDTree using params: trees=%d\n",kdtree_params.trees);
KDTreeIndex<ELEM_TYPE> kdtree(sampledDataset, kdtree_params);
t.start();
......@@ -276,7 +276,7 @@ private:
cost.searchTimeCost = searchTime;
cost.buildTimeCost = buildTime;
cost.timeCost = (buildTime*index_params.build_weight+searchTime);
logger.info("KDTree buildTime=%g, searchTime=%g, timeCost=%g\n",buildTime, searchTime, cost.timeCost);
logger().info("KDTree buildTime=%g, searchTime=%g, timeCost=%g\n",buildTime, searchTime, cost.timeCost);
}
......@@ -330,7 +330,7 @@ private:
KMeansCostData optimizeKMeans()
{
logger.info("KMEANS, Step 1: Exploring parameter space\n");
logger().info("KMEANS, Step 1: Exploring parameter space\n");
// explore kmeans parameters space using combinations of the parameters below
int maxIterations[] = { 1, 5, 10, 15 };
......@@ -363,7 +363,7 @@ private:
}
}
// logger.info("KMEANS, Step 2: simplex-downhill optimization\n");
// logger().info("KMEANS, Step 2: simplex-downhill optimization\n");
//
// const int n = 2;
// // choose initial simplex points as the best parameters so far
......@@ -397,7 +397,7 @@ private:
}
// display the costs obtained
for (int i=0;i<kmeansParamSpaceSize;++i) {
logger.info("KMeans, branching=%d, iterations=%d, time_cost=%g[%g] (build=%g, search=%g), memory_cost=%g, cost=%g\n",
logger().info("KMeans, branching=%d, iterations=%d, time_cost=%g[%g] (build=%g, search=%g), memory_cost=%g, cost=%g\n",
kmeansCosts[i].second.branching, kmeansCosts[i].second.iterations,
kmeansCosts[i].first.timeCost,kmeansCosts[i].first.timeCost/optTimeCost,
kmeansCosts[i].first.buildTimeCost, kmeansCosts[i].first.searchTimeCost,
......@@ -411,7 +411,7 @@ private:
KDTreeCostData optimizeKDTree()
{
logger.info("KD-TREE, Step 1: Exploring parameter space\n");
logger().info("KD-TREE, Step 1: Exploring parameter space\n");
// explore kd-tree parameters space using the parameters below
int testTrees[] = { 1, 4, 8, 16, 32 };
......@@ -435,7 +435,7 @@ private:
++cnt;
}
// logger.info("KD-TREE, Step 2: simplex-downhill optimization\n");
// logger().info("KD-TREE, Step 2: simplex-downhill optimization\n");
//
// const int n = 1;
// // choose initial simplex points as the best parameters so far
......@@ -467,7 +467,7 @@ private:
}
// display costs obtained
for (size_t i=0;i<kdtreeParamSpaceSize;++i) {
logger.info("kd-tree, trees=%d, time_cost=%g[%g] (build=%g, search=%g), memory_cost=%g, cost=%g\n",
logger().info("kd-tree, trees=%d, time_cost=%g[%g] (build=%g, search=%g), memory_cost=%g, cost=%g\n",
kdtreeCosts[i].second.trees,kdtreeCosts[i].first.timeCost,kdtreeCosts[i].first.timeCost/optTimeCost,
kdtreeCosts[i].first.buildTimeCost, kdtreeCosts[i].first.searchTimeCost,
kdtreeCosts[i].first.memoryCost,kdtreeCosts[i].first.totalCost);
......@@ -486,12 +486,12 @@ private:
int sampleSize = int(index_params.sample_fraction*dataset.rows);
int testSampleSize = min(sampleSize/10, 1000);
logger.info("Entering autotuning, dataset size: %d, sampleSize: %d, testSampleSize: %d\n",dataset.rows, sampleSize, testSampleSize);
logger().info("Entering autotuning, dataset size: %d, sampleSize: %d, testSampleSize: %d\n",dataset.rows, sampleSize, testSampleSize);
// For a very small dataset, it makes no sense to build any fancy index, just
// use linear search
if (testSampleSize<10) {
logger.info("Choosing linear, dataset too small\n");
logger().info("Choosing linear, dataset too small\n");
return new LinearIndexParams();
}
......@@ -501,7 +501,7 @@ private:
testDataset = random_sample(sampledDataset,testSampleSize,true);
// We compute the ground truth using linear search
logger.info("Computing ground truth... \n");
logger().info("Computing ground truth... \n");
gt_matches = Matrix<int>(new int[testDataset.rows],testDataset.rows, 1);
StartStopTimer t;
t.start();
......@@ -511,7 +511,7 @@ private:
IndexParams* bestParams = new LinearIndexParams();
// Start parameter autotune process
logger.info("Autotuning parameters...\n");
logger().info("Autotuning parameters...\n");
KMeansCostData kmeansCost = optimizeKMeans();
......@@ -554,7 +554,7 @@ private:
if (samples>0) {
Matrix<ELEM_TYPE> testDataset = random_sample(dataset,samples);
logger.info("Computing ground truth\n");
logger().info("Computing ground truth\n");
// we need to compute the ground truth first
Matrix<int> gt_matches(new int[testDataset.rows],testDataset.rows,1);
......@@ -565,12 +565,12 @@ private:
float linear = t.value;
int checks;
logger.info("Estimating number of checks\n");
logger().info("Estimating number of checks\n");
float searchTime;
float cb_index;
if (bestIndex->getType() == KMEANS) {
logger.info("KMeans algorithm, estimating cluster border factor\n");
logger().info("KMeans algorithm, estimating cluster border factor\n");
KMeansIndex<ELEM_TYPE>* kmeans = (KMeansIndex<ELEM_TYPE>*)bestIndex;
float bestSearchTime = -1;
float best_cb_index = -1;
......@@ -589,14 +589,14 @@ private:
checks = best_checks;
kmeans->set_cb_index(best_cb_index);
logger.info("Optimum cb_index: %g\n",cb_index);
logger().info("Optimum cb_index: %g\n",cb_index);
((KMeansIndexParams*)bestParams)->cb_index = cb_index;
}
else {
searchTime = test_index_precision(*bestIndex, dataset, testDataset, gt_matches, index_params.target_precision, checks, nn, 1);
}
logger.info("Required number of checks: %d \n",checks);;
logger().info("Required number of checks: %d \n",checks);;
searchParams.checks = checks;
speedup = linear/searchTime;
......
modules/flann/include/opencv2/flann/composite_index.h
View file @ f678c8f0
......@@ -58,12 +58,12 @@ struct CompositeIndexParams : public IndexParams {
void print() const
{
logger.info("Index type: %d\n",(int)algorithm);
logger.info("Trees: %d\n", trees);
logger.info("Branching: %d\n", branching);
logger.info("Iterations: %d\n", iterations);
logger.info("Centres initialisation: %d\n", centers_init);
logger.info("Cluster boundary weight: %g\n", cb_index);
logger().info("Index type: %d\n",(int)algorithm);
logger().info("Trees: %d\n", trees);
logger().info("Branching: %d\n", branching);
logger().info("Iterations: %d\n", iterations);
logger().info("Centres initialisation: %d\n", centers_init);
logger().info("Cluster boundary weight: %g\n", cb_index);
}
};
......@@ -124,9 +124,9 @@ public:
void buildIndex()
{
logger.info("Building kmeans tree...\n");
logger().info("Building kmeans tree...\n");
kmeans->buildIndex();
logger.info("Building kdtree tree...\n");
logger().info("Building kdtree tree...\n");
kdtree->buildIndex();
}
......
modules/flann/include/opencv2/flann/dist.h
View file @ f678c8f0
......@@ -82,7 +82,7 @@ double euclidean_dist(Iterator1 first1, Iterator1 last1, Iterator2 first2, doubl
return distsq;
}
double euclidean_dist(const unsigned char* first1, const unsigned char* last1, unsigned char* first2, double acc);
CV_EXPORTS double euclidean_dist(const unsigned char* first1, const unsigned char* last1, unsigned char* first2, double acc);
/**
......@@ -117,7 +117,7 @@ double manhattan_dist(Iterator1 first1, Iterator1 last1, Iterator2 first2, doubl
}
extern int flann_minkowski_order;
CV_EXPORTS int flann_minkowski_order();
/**
* Compute the Minkowski (L_p) distance between two vectors.
*
......@@ -134,7 +134,7 @@ double minkowski_dist(Iterator1 first1, Iterator1 last1, Iterator2 first2, doubl
double diff0, diff1, diff2, diff3;
Iterator1 lastgroup = last1 - 3;
int p = flann_minkowski_order;
int p = flann_minkowski_order();
/* Process 4 items with each loop for efficiency. */
while (first1 < lastgroup) {
......@@ -293,7 +293,7 @@ double kl_divergence(Iterator1 first1, Iterator1 last1, Iterator2 first2, double
extern flann_distance_t flann_distance_type;
CV_EXPORTS flann_distance_t flann_distance_type();
/**
* Custom distance function. The distance computed is dependent on the value
* of the 'flann_distance_type' global variable.
......@@ -304,7 +304,7 @@ extern flann_distance_t flann_distance_type;
template <typename Iterator1, typename Iterator2>
double custom_dist(Iterator1 first1, Iterator1 last1, Iterator2 first2, double acc = 0)
{
switch (flann_distance_type) {
switch (flann_distance_type()) {
case EUCLIDEAN:
return euclidean_dist(first1, last1, first2, acc);
case MANHATTAN:
......@@ -353,7 +353,8 @@ struct ZeroIterator {
}
};
extern ZeroIterator<float> zero;
CV_EXPORTS ZeroIterator<float>& zero();
} // namespace cvflann
......
modules/flann/include/opencv2/flann/flann.hpp
View file @ f678c8f0
......@@ -74,7 +74,7 @@ using ::cvflann::SearchParams;
template <typename T>
class Index_ {
class CV_EXPORTS Index_ {
::cvflann::Index<T>* nnIndex;
public:
......
modules/flann/include/opencv2/flann/flann_base.hpp
View file @ f678c8f0
......@@ -55,7 +55,7 @@ Sets the log level used for all flann functions
Params:
level = verbosity level
*/
void log_verbosity(int level);
CV_EXPORTS void log_verbosity(int level);
/**
......@@ -63,10 +63,10 @@ void log_verbosity(int level);
* If distance type specified is MINKOWSKI, the second argument
* specifies which order the minkowski distance should have.
*/
void set_distance_type(flann_distance_t distance_type, int order);
CV_EXPORTS void set_distance_type(flann_distance_t distance_type, int order);
struct SavedIndexParams : public IndexParams {
struct CV_EXPORTS SavedIndexParams : public IndexParams {
SavedIndexParams(std::string filename_) : IndexParams(SAVED), filename(filename_) {}
std::string filename; // filename of the stored index
......@@ -75,13 +75,13 @@ struct SavedIndexParams : public IndexParams {
void print() const
{
logger.info("Index type: %d\n",(int)algorithm);
logger.info("Filename: %s\n", filename.c_str());
logger().info("Index type: %d\n",(int)algorithm);
logger().info("Filename: %s\n", filename.c_str());
}
};
template<typename T>
class Index {
class CV_EXPORTS Index {
NNIndex<T>* nnIndex;
bool built;
......@@ -178,7 +178,7 @@ void Index<T>::knnSearch(const Matrix<T>& queries, Matrix<int>& indices, Matrix<
for (size_t i = 0; i < queries.rows; i++) {
T* target = queries[i];
resultSet.init(target, queries.cols);
resultSet.init(target, (int)queries.cols);
nnIndex->findNeighbors(resultSet, target, searchParams);
......@@ -202,7 +202,7 @@ int Index<T>::radiusSearch(const Matrix<T>& query, Matrix<int>& indices, Matrix<
assert(query.cols==nnIndex->veclen());
RadiusResultSet<T> resultSet(radius);
resultSet.init(query.data, query.cols);
resultSet.init(query.data, (int)query.cols);
nnIndex->findNeighbors(resultSet,query.data,searchParams);
// TODO: optimise here
......@@ -217,7 +217,7 @@ int Index<T>::radiusSearch(const Matrix<T>& query, Matrix<int>& indices, Matrix<
dists[0][i] = distances[i];
}
return count_nn;
return (int)count_nn;
}
......
modules/flann/include/opencv2/flann/general.h
View file @ f678c8f0
......@@ -31,9 +31,16 @@
#ifndef CONSTANTS_H
#define CONSTANTS_H
#define ARRAY_LEN(a) (sizeof(a)/sizeof(a[0]))
#ifdef __cplusplus
#include <stdexcept>
#include <cassert>
#include "opencv2/flann/object_factory.h"
namespace cvflann {
#undef ARRAY_LEN
#define ARRAY_LEN(a) (sizeof(a)/sizeof(a[0]))
/* Nearest neighbour index algorithms */
enum flann_algorithm_t {
......@@ -85,17 +92,6 @@ enum flann_datatype_t {
FLOAT64 = 9
};
#ifdef __cplusplus
#include <stdexcept>
#include <cassert>
#include "opencv2/flann/object_factory.h"
namespace cvflann {
template <typename ELEM_TYPE>
struct DistType
{
......@@ -123,7 +119,7 @@ class FLANNException : public std::runtime_error {
};
struct IndexParams {
struct CV_EXPORTS IndexParams {
protected:
IndexParams(flann_algorithm_t algorithm_) : algorithm(algorithm_) {};
......@@ -139,7 +135,7 @@ public:
typedef ObjectFactory<IndexParams, flann_algorithm_t> ParamsFactory;
struct SearchParams {
struct CV_EXPORTS SearchParams {
SearchParams(int checks_ = 32) :
checks(checks_) {};
......
modules/flann/include/opencv2/flann/index_testing.h
View file @ f678c8f0
......@@ -46,7 +46,7 @@ using namespace std;
namespace cvflann
{
int countCorrectMatches(int* neighbors, int* groundTruth, int n);
CV_EXPORTS int countCorrectMatches(int* neighbors, int* groundTruth, int n);
template <typename ELEM_TYPE>
......@@ -73,7 +73,7 @@ template <typename ELEM_TYPE>
float search_with_ground_truth(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>& inputData, const Matrix<ELEM_TYPE>& testData, const Matrix<int>& matches, int nn, int checks, float& time, float& dist, int skipMatches)
{
if (matches.cols<size_t(nn)) {
logger.info("matches.cols=%d, nn=%d\n",matches.cols,nn);
logger().info("matches.cols=%d, nn=%d\n",matches.cols,nn);
throw FLANNException("Ground truth is not computed for as many neighbors as requested");
}
......@@ -109,7 +109,7 @@ float search_with_ground_truth(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE
dist = distR/(testData.rows*nn);
logger.info("%8d %10.4g %10.5g %10.5g %10.5g\n",
logger().info("%8d %10.4g %10.5g %10.5g %10.5g\n",
checks, precicion, time, 1000.0 * time / testData.rows, dist);
return precicion;
......@@ -120,8 +120,8 @@ template <typename ELEM_TYPE>
float test_index_checks(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>& inputData, const Matrix<ELEM_TYPE>& testData, const Matrix<int>& matches,
int checks, float& precision, int nn = 1, int skipMatches = 0)
{
logger.info(" Nodes Precision(%) Time(s) Time/vec(ms) Mean dist\n");
logger.info("---------------------------------------------------------\n");
logger().info(" Nodes Precision(%) Time(s) Time/vec(ms) Mean dist\n");
logger().info("---------------------------------------------------------\n");
float time = 0;
float dist = 0;
......@@ -136,8 +136,8 @@ float test_index_precision(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>& i
{
const float SEARCH_EPS = 0.001;
logger.info(" Nodes Precision(%) Time(s) Time/vec(ms) Mean dist\n");
logger.info("---------------------------------------------------------\n");
logger().info(" Nodes Precision(%) Time(s) Time/vec(ms) Mean dist\n");
logger().info("---------------------------------------------------------\n");
int c2 = 1;
float p2;
......@@ -149,7 +149,7 @@ float test_index_precision(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>& i
p2 = search_with_ground_truth(index, inputData, testData, matches, nn, c2, time, dist, skipMatches);
if (p2>precision) {
logger.info("Got as close as I can\n");
logger().info("Got as close as I can\n");
checks = c2;
return time;
}
......@@ -164,7 +164,7 @@ float test_index_precision(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>& i
int cx;
float realPrecision;
if (fabs(p2-precision)>SEARCH_EPS) {
logger.info("Start linear estimation\n");
logger().info("Start linear estimation\n");
// after we got to values in the vecinity of the desired precision
// use linear approximation get a better estimation
......@@ -180,7 +180,7 @@ float test_index_precision(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>& i
}
cx = (c1+c2)/2;
if (cx==c1) {
logger.info("Got as close as I can\n");
logger().info("Got as close as I can\n");
break;
}
realPrecision = search_with_ground_truth(index, inputData, testData, matches, nn, cx, time, dist, skipMatches);
......@@ -190,7 +190,7 @@ float test_index_precision(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>& i
p2 = realPrecision;
} else {
logger.info("No need for linear estimation\n");
logger().info("No need for linear estimation\n");
cx = c2;
realPrecision = p2;
}
......@@ -212,8 +212,8 @@ float test_index_precisions(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>&
int pindex = 0;
float precision = precisions[pindex];
logger.info(" Nodes Precision(%) Time(s) Time/vec(ms) Mean dist");
logger.info("---------------------------------------------------------");
logger().info(" Nodes Precision(%) Time(s) Time/vec(ms) Mean dist");
logger().info("---------------------------------------------------------");
int c2 = 1;
float p2;
......@@ -234,7 +234,7 @@ float test_index_precisions(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>&
}
if (pindex==precisions_length) {
logger.info("Got as close as I can\n");
logger().info("Got as close as I can\n");
return time;
}
......@@ -252,7 +252,7 @@ float test_index_precisions(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>&
int cx;
float realPrecision;
if (fabs(p2-precision)>SEARCH_EPS) {
logger.info("Start linear estimation\n");
logger().info("Start linear estimation\n");
// after we got to values in the vecinity of the desired precision
// use linear approximation get a better estimation
......@@ -268,7 +268,7 @@ float test_index_precisions(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>&
}
cx = (c1+c2)/2;
if (cx==c1) {
logger.info("Got as close as I can\n");
logger().info("Got as close as I can\n");
break;
}
realPrecision = search_with_ground_truth(index, inputData, testData, matches, nn, cx, time, dist, skipMatches);
......@@ -278,7 +278,7 @@ float test_index_precisions(NNIndex<ELEM_TYPE>& index, const Matrix<ELEM_TYPE>&
p2 = realPrecision;
} else {
logger.info("No need for linear estimation\n");
logger().info("No need for linear estimation\n");
cx = c2;
realPrecision = p2;
}
......
modules/flann/include/opencv2/flann/kdtree_index.h
View file @ f678c8f0
......@@ -51,7 +51,7 @@ using namespace std;
namespace cvflann
{
struct KDTreeIndexParams : public IndexParams {
struct CV_EXPORTS KDTreeIndexParams : public IndexParams {
KDTreeIndexParams(int trees_ = 4) : IndexParams(KDTREE), trees(trees_) {};
int trees; // number of randomized trees to use (for kdtree)
......@@ -60,8 +60,8 @@ struct KDTreeIndexParams : public IndexParams {
void print() const
{
logger.info("Index type: %d\n",(int)algorithm);
logger.info("Trees: %d\n", trees);
logger().info("Index type: %d\n",(int)algorithm);
logger().info("Trees: %d\n", trees);
}
};
......
modules/flann/include/opencv2/flann/kmeans_index.h
View file @ f678c8f0
......@@ -52,7 +52,7 @@ namespace cvflann
{
struct KMeansIndexParams : public IndexParams {
struct CV_EXPORTS KMeansIndexParams : public IndexParams {
KMeansIndexParams(int branching_ = 32, int iterations_ = 11,
flann_centers_init_t centers_init_ = CENTERS_RANDOM, float cb_index_ = 0.2 ) :
IndexParams(KMEANS),
......@@ -70,11 +70,11 @@ struct KMeansIndexParams : public IndexParams {
void print() const
{
logger.info("Index type: %d\n",(int)algorithm);
logger.info("Branching: %d\n", branching);
logger.info("Iterations: %d\n", iterations);
logger.info("Centres initialisation: %d\n", centers_init);
logger.info("Cluster boundary weight: %g\n", cb_index);
logger().info("Index type: %d\n",(int)algorithm);
logger().info("Branching: %d\n", branching);
logger().info("Iterations: %d\n", iterations);
logger().info("Centres initialisation: %d\n", centers_init);
logger().info("Cluster boundary weight: %g\n", cb_index);
}
};
......@@ -578,7 +578,7 @@ public:
int clusterCount = getMinVarianceClusters(root, clusters, numClusters, variance);
// logger.info("Clusters requested: %d, returning %d\n",numClusters, clusterCount);
// logger().info("Clusters requested: %d, returning %d\n",numClusters, clusterCount);
for (int i=0;i<clusterCount;++i) {
......@@ -671,13 +671,13 @@ private:
for (size_t j=0;j<veclen_;++j) {
mean[j] += vec[j];
}
variance += flann_dist(vec,vec+veclen_,zero);
variance += flann_dist(vec,vec+veclen_,zero());
}
for (size_t j=0;j<veclen_;++j) {
mean[j] /= size_;
}
variance /= size_;
variance -= flann_dist(mean,mean+veclen_,zero);
variance -= flann_dist(mean,mean+veclen_,zero());
DIST_TYPE tmp = 0;
for (int i=0;i<indices_length;++i) {
......@@ -856,7 +856,7 @@ private:
float mean_radius =0;
for (int i=0;i<indices_length;++i) {
if (belongs_to[i]==c) {
float d = flann_dist(dataset[indices[i]],dataset[indices[i]]+veclen_,zero);
float d = flann_dist(dataset[indices[i]],dataset[indices[i]]+veclen_,zero());
variance += d;
mean_radius += sqrt(d);
swap(indices[i],indices[end]);
......@@ -866,7 +866,7 @@ private:
}
variance /= s;
mean_radius /= s;
variance -= flann_dist(centers[c],centers[c]+veclen_,zero);
variance -= flann_dist(centers[c],centers[c]+veclen_,zero());
node->childs[c] = pool.allocate<KMeansNodeSt>();
node->childs[c]->radius = radiuses[c];
......
modules/flann/include/opencv2/flann/linear_index.h
View file @ f678c8f0
......@@ -38,14 +38,14 @@
namespace cvflann
{
struct LinearIndexParams : public IndexParams {
struct CV_EXPORTS LinearIndexParams : public IndexParams {
LinearIndexParams() : IndexParams(LINEAR) {};
flann_algorithm_t getIndexType() const { return algorithm; }
void print() const
{
logger.info("Index type: %d\n",(int)algorithm);
logger().info("Index type: %d\n",(int)algorithm);
}
};
......
modules/flann/include/opencv2/flann/logger.h
View file @ f678c8f0
......@@ -40,7 +40,7 @@ using namespace std;
namespace cvflann
{
class Logger
class CV_EXPORTS Logger
{
FILE* stream;
int logLevel;
......@@ -84,7 +84,7 @@ public:
int info(const char* fmt, ...);
};
extern Logger logger;
CV_EXPORTS Logger& logger();
} // namespace cvflann
......
modules/flann/include/opencv2/flann/object_factory.h
View file @ f678c8f0
......@@ -31,6 +31,7 @@
#ifndef OBJECT_FACTORY_H_
#define OBJECT_FACTORY_H_
#include "opencv2/core/types_c.h"
#include <map>
namespace cvflann
......
modules/flann/include/opencv2/flann/random.h
View file @ f678c8f0
......@@ -43,17 +43,17 @@ namespace cvflann
/**
* Seeds the random number generator
*/
void seed_random(unsigned int seed);
CV_EXPORTS void seed_random(unsigned int seed);
/*
* Generates a random double value.
*/
double rand_double(double high = 1.0, double low=0);
CV_EXPORTS double rand_double(double high = 1.0, double low=0);
/*
* Generates a random integer value.
*/
int rand_int(int high = RAND_MAX, int low = 0);
CV_EXPORTS int rand_int(int high = RAND_MAX, int low = 0);
/**
......@@ -63,7 +63,7 @@ int rand_int(int high = RAND_MAX, int low = 0);
* TODO: improve on this to use a generator function instead of an
* array of randomly permuted numbers
*/
class UniqueRandom
class CV_EXPORTS UniqueRandom
{
int* vals;
int size;
......
modules/flann/include/opencv2/flann/result_set.h
View file @ f678c8f0
......@@ -161,7 +161,7 @@ public:
for (int i=0;i<count;++i) {
if (indices[i]==index) return false;
}
float dist = flann_dist(target, target_end, point);
float dist = (float)flann_dist(target, target_end, point);
if (count<capacity) {
indices[count] = index;
......
modules/flann/include/opencv2/flann/saving.h
View file @ f678c8f0
......@@ -47,13 +47,13 @@ template<> struct Datatype<float> { static flann_datatype_t type() { return FLOA
template<> struct Datatype<double> { static flann_datatype_t type() { return FLOAT64; } };
extern const char FLANN_SIGNATURE[];
extern const char FLANN_VERSION[];
CV_EXPORTS const char* FLANN_SIGNATURE();
CV_EXPORTS const char* FLANN_VERSION();
/**
* Structure representing the index header.
*/
struct IndexHeader
struct CV_EXPORTS IndexHeader
{
char signature[16];
char version[16];
......@@ -74,12 +74,12 @@ void save_header(FILE* stream, const NNIndex<ELEM_TYPE>& index)
{
IndexHeader header;
memset(header.signature, 0 , sizeof(header.signature));
strcpy(header.signature, FLANN_SIGNATURE);
strcpy(header.signature, FLANN_SIGNATURE());
memset(header.version, 0 , sizeof(header.version));
strcpy(header.version, FLANN_VERSION);
strcpy(header.version, FLANN_VERSION());
header.data_type = Datatype<ELEM_TYPE>::type();
header.index_type = index.getType();
header.rows = index.size();
header.rows = (int)index.size();
header.cols = index.veclen();
std::fwrite(&header, sizeof(header),1,stream);
......@@ -91,7 +91,7 @@ void save_header(FILE* stream, const NNIndex<ELEM_TYPE>& index)
* @param stream - Stream to load from
* @return Index header
*/
IndexHeader load_header(FILE* stream);
CV_EXPORTS IndexHeader load_header(FILE* stream);
template<typename T>
......
modules/flann/include/opencv2/flann/simplex_downhill.h
View file @ f678c8f0
......@@ -123,7 +123,7 @@ float optimizeSimplexDownhill(T* points, int n, F func, float* vals = NULL )
if (val_r>=vals[0] && val_r<vals[n]) {
// reflection between second highest and lowest
// add it to the simplex
logger.info("Choosing reflection\n");
logger().info("Choosing reflection\n");
addValue(n, val_r,vals, p_r, points, n);
continue;
}
......@@ -138,11 +138,11 @@ float optimizeSimplexDownhill(T* points, int n, F func, float* vals = NULL )
float val_e = func(p_e);
if (val_e<val_r) {
logger.info("Choosing reflection and expansion\n");
logger().info("Choosing reflection and expansion\n");
addValue(n, val_e,vals,p_e,points,n);
}
else {
logger.info("Choosing reflection\n");
logger().info("Choosing reflection\n");
addValue(n, val_r,vals,p_r,points,n);
}
continue;
......@@ -154,13 +154,13 @@ float optimizeSimplexDownhill(T* points, int n, F func, float* vals = NULL )
float val_e = func(p_e);
if (val_e<vals[n]) {
logger.info("Choosing contraction\n");
logger().info("Choosing contraction\n");
addValue(n,val_e,vals,p_e,points,n);
continue;
}
}
{
logger.info("Full contraction\n");
logger().info("Full contraction\n");
for (int j=1;j<=n;++j) {
for (int i=0;i<n;++i) {
points[j*n+i] = (points[j*n+i]+points[i])/2;
......
modules/flann/include/opencv2/flann/timer.h
View file @ f678c8f0
......@@ -42,7 +42,7 @@ namespace cvflann
*
* Can be used to time portions of code.
*/
class StartStopTimer
class CV_EXPORTS StartStopTimer
{
clock_t startTime;
......
modules/flann/src/flann.cpp
View file @ f678c8f0
......@@ -26,19 +26,7 @@
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*************************************************************************/
#include <cstdio>
#include <cstdarg>
#include <sstream>
#include "opencv2/flann/dist.h"
#include "opencv2/flann/index_testing.h"
#include "opencv2/flann/logger.h"
#include "opencv2/flann/logger.h"
#include "opencv2/flann/saving.h"
#include "opencv2/flann/general.h"
// index types
#include "opencv2/flann/all_indices.h"
#include "precomp.hpp"
namespace cvflann
{
......@@ -47,17 +35,20 @@ namespace cvflann
/** Global variable indicating the distance metric
* to be used.
*/
flann_distance_t flann_distance_type = EUCLIDEAN;
flann_distance_t flann_distance_type_ = EUCLIDEAN;
flann_distance_t flann_distance_type() { return flann_distance_type_; }
/**
* Zero iterator that emulates a zero feature.
*/
ZeroIterator<float> zero;
ZeroIterator<float> zero_;
ZeroIterator<float>& zero() { return zero_; }
/**
* Order of Minkowski distance to use.
*/
int flann_minkowski_order;
int flann_minkowski_order_;
int flann_minkowski_order() { return flann_minkowski_order_; }
double euclidean_dist(const unsigned char* first1, const unsigned char* last1, unsigned char* first2, double acc)
......@@ -98,9 +89,11 @@ int countCorrectMatches(int* neighbors, int* groundTruth, int n)
return count;
}
// ----------------------- logger.cpp ---------------------------
// ----------------------- logger().cpp ---------------------------
Logger logger_;
Logger logger;
Logger& logger() { return logger_; }
int Logger::log(int level, const char* fmt, ...)
{
......@@ -163,20 +156,22 @@ int rand_int(int high, int low)
// ----------------------- saving.cpp ---------------------------
const char FLANN_SIGNATURE[] = "FLANN_INDEX";
const char FLANN_VERSION[] = "1.5.0";
const char FLANN_SIGNATURE_[] = "FLANN_INDEX";
const char FLANN_VERSION_[] = "1.5.0";
const char* FLANN_SIGNATURE() { return FLANN_SIGNATURE_; }
const char* FLANN_VERSION() { return FLANN_VERSION_; }
IndexHeader load_header(FILE* stream)
{
IndexHeader header;
int read_size = fread(&header,sizeof(header),1,stream);
size_t read_size = fread(&header,sizeof(header),1,stream);
if (read_size!=1) {
throw FLANNException("Invalid index file, cannot read");
}
if (strcmp(header.signature,FLANN_SIGNATURE)!=0) {
if (strcmp(header.signature,FLANN_SIGNATURE())!=0) {
throw FLANNException("Invalid index file, wrong signature");
}
......@@ -190,14 +185,14 @@ IndexHeader load_header(FILE* stream)
void log_verbosity(int level)
{
if (level>=0) {
logger.setLevel(level);
logger().setLevel(level);
}
}
void set_distance_type(flann_distance_t distance_type, int order)
{
flann_distance_type = distance_type;
flann_minkowski_order = order;
flann_distance_type_ = distance_type;
flann_minkowski_order_ = order;
}
class StaticInit
......
modules/flann/src/precomp.cpp 0 → 100644
View file @ f678c8f0
#include "precomp.hpp"
modules/flann/src/precomp.hpp
View file @ f678c8f0
#ifndef _OPENCV_FLANN_PRECOMP_HPP_
#define _OPENCV_FLANN_PRECOMP_HPP_
#include <cstdio>
#include <cstdarg>
#include <sstream>
#include "opencv2/flann/dist.h"
#include "opencv2/flann/index_testing.h"
#include "opencv2/flann/logger.h"
#include "opencv2/flann/saving.h"
#include "opencv2/flann/general.h"
// index types
#include "opencv2/flann/all_indices.h"
#endif
modules/objdetect/src/_lsvm_error.h
View file @ f678c8f0
#ifndef SVM_ERROR
#define SVM_ERROR
#ifndef LSVM_ERROR
#define LSVM_ERROR
#define LATENT_SVM_OK 0
#define DISTANCE_TRANSFORM_OK 1
......
samples/c/find_obj.cpp
View file @ f678c8f0
......@@ -171,7 +171,7 @@ locatePlanarObject( const CvSeq* objectKeypoints, const CvSeq* objectDescriptors
findPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs );
#endif
n = ptpairs.size()/2;
n = (int)(ptpairs.size()/2);
if( n < 4 )
return 0;
......
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