void ImageMolecule::deserialize(const cv::FileNode& fn)
{
FileNode atoms = fn["atoms"];
CV_Assert(atoms.type() == FileNode::SEQ);
std::map<int, Ptr<ImageAtom> > a_map;
for (size_t i = 0; i < atoms.size(); i++)
{
Ptr<ImageAtom> atom(new ImageAtom);
atom->deserialize(atoms[i]);
a_map[atom->uid()] = atom;
//we will insert from pairs...
insertAtom(atom);
}
FileNode pairs = fn["pairs"];
CV_Assert(pairs.type() == FileNode::SEQ);
vector<AtomPair> pairs_temp;
pairs_temp.resize(pairs.size());
for (size_t i = 0; i < pairs.size(); i++)
{
pairs_temp[i].deserialize(pairs[i]);
pairs_temp[i].setAtom1(a_map[pairs_temp[i].atom1()->uid()]);
pairs_temp[i].setAtom2(a_map[pairs_temp[i].atom2()->uid()]);
}
insertPairs(pairs_temp);
}
void CvCascadeBoostTree::read( const FileNode &node, CvBoost* _ensemble,
CvDTreeTrainData* _data )
{
int maxCatCount = ((CvCascadeBoostTrainData*)_data)->featureEvaluator->getMaxCatCount();
int subsetN = (maxCatCount + 31)/32;
int step = 3 + ( maxCatCount>0 ? subsetN : 1 );
queue<CvDTreeNode*> internalNodesQueue;
FileNodeIterator internalNodesIt, leafValsuesIt;
CvDTreeNode* prntNode, *cldNode;
clear();
data = _data;
ensemble = _ensemble;
pruned_tree_idx = 0;
// read tree nodes
FileNode rnode = node[CC_INTERNAL_NODES];
internalNodesIt = rnode.end();
leafValsuesIt = node[CC_LEAF_VALUES].end();
internalNodesIt--; leafValsuesIt--;
for( size_t i = 0; i < rnode.size()/step; i++ )
{
prntNode = data->new_node( 0, 0, 0, 0 );
if ( maxCatCount > 0 )
{
prntNode->split = data->new_split_cat( 0, 0 );
for( int j = subsetN-1; j>=0; j--)
{
*internalNodesIt >> prntNode->split->subset[j]; internalNodesIt--;
}
}
else
{
void ReadRightRects(vector<ImageRecognition::SlidingRect> &rightRects, const string &xml_filename, RecognitionStatistics &stat)
{
using namespace Utils;
FileStorage file_storage(xml_filename, FileStorage::READ);
FileNode images = file_storage["images"];
rightRects.reserve(images.size());
for (FileNodeIterator it = images.begin(); it != images.end(); ++it)
{
string part_filename = string(*it);
int dot_pos = part_filename.find_first_of('.');
if (dot_pos != -1)
part_filename = part_filename.substr(0, dot_pos);
stringstream ss(part_filename);
vector<string> parts;
string part;
while (getline(ss, part, '_'))
parts.push_back(part);
rightRects.push_back(ImageRecognition::SlidingRect());
int last = parts.size() - 1;
rightRects.back().rect.x = str2int(parts[last - 3]);
rightRects.back().rect.y = str2int(parts[last - 2]);
rightRects.back().rect.width = str2int(parts[last - 1]);
rightRects.back().rect.height = str2int(parts[last]);
}
}
void loadHist(mH2& hist){
FileStorage fs("test123.xml", FileStorage::READ);
FileNode n = fs["ModelHistograms"];
// Loop through Classes
for(int i=0;i<n.size();i++){
stringstream ss;
ss << "Class_";
ss << i;
string a = ss.str();
FileNode n1 = n[a];
// Loop through Each classes Models
for(int j = 0; j < n1.size(); j++){
stringstream ss1;
ss1 << "Model_";
ss1 << j;
string b = ss1.str();
FileNode n2 = n1[b];
// Save stored Mat to mask
FileNodeIterator it = n2.begin(), it_end = n2.end();
for(;it != it_end;++it){
Mat mask;
(*it) >> hist[i][j];
}
}
}
fs.release();
}
void Expression::load(std::string filename) {
FileStorage fs(ofToDataPath(filename), FileStorage::READ);
description = (std::string) fs["description"];
FileNode samplesNode = fs["samples"];
int n = samplesNode.size();
samples.resize(n);
for(int i = 0; i < n; i++) {
samplesNode[i] >> samples[i];
}
}
void NMPTUtils::readMatBinary(const FileNode &tm, Mat &mat) {
//FileNode tm = fs[name];
int rows = (int)tm["rows"], cols = (int)tm["cols"], type = (int)tm["type"];
mat.create(rows,cols,type);
if (rows > 0 && cols > 0) {
vector<string> vs;
FileNode tl = tm["data"];
//std::cout << tl.type() << std::endl;
CV_Assert(tl.type() == FileNode::SEQ);
vs.resize(tl.size());
for (size_t i = 0; i < tl.size(); i++) {
tl[i] >> vs[i];
}
// CV_Assert(tl.size() == (size_t)numRegs);
// tm["data"] >> vs;
string s;
joinString(vs, s);
asciiToBinary(s, mat.data, mat.rows*mat.step);
}
bool HOGEvaluator::read( const FileNode& node )
{
features->resize(node.size());
featuresPtr = &(*features)[0];
FileNodeIterator it = node.begin(), it_end = node.end();
for(int i = 0; it != it_end; ++it, i++)
{
if(!featuresPtr[i].read(*it))
return false;
}
return true;
}
bool CvCascadeClassifier::readStages( const FileNode &node)
{
FileNode rnode = node[CC_STAGES];
if (!rnode.empty() || !rnode.isSeq())
return false;
stageClassifiers.reserve(numStages);
FileNodeIterator it = rnode.begin();
for( int i = 0; i < min( (int)rnode.size(), numStages ); i++, it++ )
{
Ptr<CvCascadeBoost> tempStage = makePtr<CvCascadeBoost>();
if ( !tempStage->read( *it, featureEvaluator, *stageParams) )
return false;
stageClassifiers.push_back(tempStage);
}
return true;
}
virtual int readRunParams( FileStorage& fs )
{
int code = CV_StereoMatchingTest::readRunParams(fs);
FileNode fn = fs.getFirstTopLevelNode();
assert(fn.isSeq());
for( int i = 0; i < (int)fn.size(); i+=4 )
{
string caseName = fn[i], datasetName = fn[i+1];
RunParams params;
string ndisp = fn[i+2]; params.ndisp = atoi(ndisp.c_str());
string iterCount = fn[i+3]; params.iterCount = atoi(iterCount.c_str());
caseNames.push_back( caseName );
caseDatasets.push_back( datasetName );
caseRunParams.push_back( params );
}
return code;
}
virtual int readRunParams( FileStorage& fs )
{
int code = CV_StereoMatchingTest::readRunParams(fs);
FileNode fn = fs.getFirstTopLevelNode();
assert(fn.isSeq());
for( int i = 0; i < (int)fn.size(); i+=5 )
{
string caseName = fn[i], datasetName = fn[i+1];
RunParams params;
string ndisp = fn[i+2]; params.ndisp = atoi(ndisp.c_str());
string winSize = fn[i+3]; params.winSize = atoi(winSize.c_str());
string fullDP = fn[i+4]; params.fullDP = atoi(fullDP.c_str()) == 0 ? false : true;
caseNames.push_back( caseName );
caseDatasets.push_back( datasetName );
caseRunParams.push_back( params );
}
return code;
}
void read(const FileNode& fn)
{
clear();
read_params(fn["training_params"]);
fn["weights"] >> weights;
fn["means"] >> means;
FileNode cfn = fn["covs"];
FileNodeIterator cfn_it = cfn.begin();
int i, n = (int)cfn.size();
covs.resize(n);
for( i = 0; i < n; i++, ++cfn_it )
(*cfn_it) >> covs[i];
decomposeCovs();
computeLogWeightDivDet();
}
bool CvCascadeClassifier::readStages( const FileNode &node)
{
FileNode rnode = node[CC_STAGES];
if (!rnode.empty() || !rnode.isSeq())
return false;
stageClassifiers.reserve(numStages);
FileNodeIterator it = rnode.begin();
for( int i = 0; i < min( (int)rnode.size(), numStages ); i++, it++ )
{
CvCascadeBoost* tempStage = new CvCascadeBoost;
if ( !tempStage->read( *it, (CvFeatureEvaluator *)featureEvaluator, *((CvCascadeBoostParams*)stageParams) ) )
{
delete tempStage;
return false;
}
stageClassifiers.push_back(tempStage);
}
return true;
}
int CV_StereoMatchingTest::readDatasetsParams( FileStorage& fs )
{
if( !fs.isOpened() )
{
ts->printf( CvTS::LOG, "datasetsParams can not be read " );
return CvTS::FAIL_INVALID_TEST_DATA;
}
datasetsParams.clear();
FileNode fn = fs.getFirstTopLevelNode();
assert(fn.isSeq());
for( int i = 0; i < (int)fn.size(); i+=3 )
{
string name = fn[i];
DatasetParams params;
string sf = fn[i+1]; params.dispScaleFactor = atoi(sf.c_str());
string uv = fn[i+2]; params.dispUnknVal = atoi(uv.c_str());
datasetsParams[name] = params;
}
return CvTS::OK;
}
int CV_MLBaseTest::read_params( CvFileStorage* __fs )
{
FileStorage _fs(__fs, false);
if( !_fs.isOpened() )
test_case_count = -1;
else
{
FileNode fn = _fs.getFirstTopLevelNode()["run_params"][modelName];
test_case_count = (int)fn.size();
if( test_case_count <= 0 )
test_case_count = -1;
if( test_case_count > 0 )
{
dataSetNames.resize( test_case_count );
FileNodeIterator it = fn.begin();
for( int i = 0; i < test_case_count; i++, ++it )
{
dataSetNames[i] = (string)*it;
}
}
}
return cvtest::TS::OK;;
}
void run(int)
{
double ranges[][2] = {{0, 256}, {-128, 128}, {0, 65536}, {-32768, 32768},
{-1000000, 1000000}, {-10, 10}, {-10, 10}};
RNG& rng = ts->get_rng();
RNG rng0;
test_case_count = 4;
int progress = 0;
MemStorage storage(cvCreateMemStorage(0));
for( int idx = 0; idx < test_case_count; idx++ )
{
ts->update_context( this, idx, false );
progress = update_progress( progress, idx, test_case_count, 0 );
cvClearMemStorage(storage);
bool mem = (idx % 4) >= 2;
string filename = tempfile(idx % 2 ? ".yml" : ".xml");
FileStorage fs(filename, FileStorage::WRITE + (mem ? FileStorage::MEMORY : 0));
int test_int = (int)cvtest::randInt(rng);
double test_real = (cvtest::randInt(rng)%2?1:-1)*exp(cvtest::randReal(rng)*18-9);
string test_string = "vw wv23424rt\"&<>&'@#$@$%$%&%IJUKYILFD@#$@%$&*&() ";
int depth = cvtest::randInt(rng) % (CV_64F+1);
int cn = cvtest::randInt(rng) % 4 + 1;
Mat test_mat(cvtest::randInt(rng)%30+1, cvtest::randInt(rng)%30+1, CV_MAKETYPE(depth, cn));
rng0.fill(test_mat, CV_RAND_UNI, Scalar::all(ranges[depth][0]), Scalar::all(ranges[depth][1]));
if( depth >= CV_32F )
{
exp(test_mat, test_mat);
Mat test_mat_scale(test_mat.size(), test_mat.type());
rng0.fill(test_mat_scale, CV_RAND_UNI, Scalar::all(-1), Scalar::all(1));
multiply(test_mat, test_mat_scale, test_mat);
}
CvSeq* seq = cvCreateSeq(test_mat.type(), (int)sizeof(CvSeq),
(int)test_mat.elemSize(), storage);
cvSeqPushMulti(seq, test_mat.data, test_mat.cols*test_mat.rows);
CvGraph* graph = cvCreateGraph( CV_ORIENTED_GRAPH,
sizeof(CvGraph), sizeof(CvGraphVtx),
sizeof(CvGraphEdge), storage );
int edges[][2] = {{0,1},{1,2},{2,0},{0,3},{3,4},{4,1}};
int i, vcount = 5, ecount = 6;
for( i = 0; i < vcount; i++ )
cvGraphAddVtx(graph);
for( i = 0; i < ecount; i++ )
{
CvGraphEdge* edge;
cvGraphAddEdge(graph, edges[i][0], edges[i][1], 0, &edge);
edge->weight = (float)(i+1);
}
depth = cvtest::randInt(rng) % (CV_64F+1);
cn = cvtest::randInt(rng) % 4 + 1;
int sz[] = {cvtest::randInt(rng)%10+1, cvtest::randInt(rng)%10+1, cvtest::randInt(rng)%10+1};
MatND test_mat_nd(3, sz, CV_MAKETYPE(depth, cn));
rng0.fill(test_mat_nd, CV_RAND_UNI, Scalar::all(ranges[depth][0]), Scalar::all(ranges[depth][1]));
if( depth >= CV_32F )
{
exp(test_mat_nd, test_mat_nd);
MatND test_mat_scale(test_mat_nd.dims, test_mat_nd.size, test_mat_nd.type());
rng0.fill(test_mat_scale, CV_RAND_UNI, Scalar::all(-1), Scalar::all(1));
multiply(test_mat_nd, test_mat_scale, test_mat_nd);
}
int ssz[] = {cvtest::randInt(rng)%10+1, cvtest::randInt(rng)%10+1,
cvtest::randInt(rng)%10+1,cvtest::randInt(rng)%10+1};
SparseMat test_sparse_mat = cvTsGetRandomSparseMat(4, ssz, cvtest::randInt(rng)%(CV_64F+1),
cvtest::randInt(rng) % 10000, 0, 100, rng);
fs << "test_int" << test_int << "test_real" << test_real << "test_string" << test_string;
fs << "test_mat" << test_mat;
fs << "test_mat_nd" << test_mat_nd;
fs << "test_sparse_mat" << test_sparse_mat;
fs << "test_list" << "[" << 0.0000000000001 << 2 << CV_PI << -3435345 << "2-502 2-029 3egegeg" <<
"{:" << "month" << 12 << "day" << 31 << "year" << 1969 << "}" << "]";
fs << "test_map" << "{" << "x" << 1 << "y" << 2 << "width" << 100 << "height" << 200 << "lbp" << "[:";
const uchar arr[] = {0, 1, 1, 0, 1, 1, 0, 1};
fs.writeRaw("u", arr, (int)(sizeof(arr)/sizeof(arr[0])));
fs << "]" << "}";
cvWriteComment(*fs, "test comment", 0);
fs.writeObj("test_seq", seq);
fs.writeObj("test_graph",graph);
CvGraph* graph2 = (CvGraph*)cvClone(graph);
string content = fs.releaseAndGetString();
if(!fs.open(mem ? content : filename, FileStorage::READ + (mem ? FileStorage::MEMORY : 0)))
{
ts->printf( cvtest::TS::LOG, "filename %s can not be read\n", !mem ? filename.c_str() : content.c_str());
ts->set_failed_test_info( cvtest::TS::FAIL_MISSING_TEST_DATA );
return;
}
int real_int = (int)fs["test_int"];
double real_real = (double)fs["test_real"];
string real_string = (string)fs["test_string"];
if( real_int != test_int ||
fabs(real_real - test_real) > DBL_EPSILON*(fabs(test_real)+1) ||
real_string != test_string )
{
ts->printf( cvtest::TS::LOG, "the read scalars are not correct\n" );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
CvMat* m = (CvMat*)fs["test_mat"].readObj();
CvMat _test_mat = test_mat;
double max_diff = 0;
CvMat stub1, _test_stub1;
cvReshape(m, &stub1, 1, 0);
cvReshape(&_test_mat, &_test_stub1, 1, 0);
vector<int> pt;
if( !m || !CV_IS_MAT(m) || m->rows != test_mat.rows || m->cols != test_mat.cols ||
cvtest::cmpEps( Mat(&stub1), Mat(&_test_stub1), &max_diff, 0, &pt, true) < 0 )
{
ts->printf( cvtest::TS::LOG, "the read matrix is not correct: (%.20g vs %.20g) at (%d,%d)\n",
cvGetReal2D(&stub1, pt[0], pt[1]), cvGetReal2D(&_test_stub1, pt[0], pt[1]),
pt[0], pt[1] );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
if( m && CV_IS_MAT(m))
cvReleaseMat(&m);
CvMatND* m_nd = (CvMatND*)fs["test_mat_nd"].readObj();
CvMatND _test_mat_nd = test_mat_nd;
if( !m_nd || !CV_IS_MATND(m_nd) )
{
ts->printf( cvtest::TS::LOG, "the read nd-matrix is not correct\n" );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
CvMat stub, _test_stub;
cvGetMat(m_nd, &stub, 0, 1);
cvGetMat(&_test_mat_nd, &_test_stub, 0, 1);
cvReshape(&stub, &stub1, 1, 0);
cvReshape(&_test_stub, &_test_stub1, 1, 0);
if( !CV_ARE_TYPES_EQ(&stub, &_test_stub) ||
!CV_ARE_SIZES_EQ(&stub, &_test_stub) ||
//cvNorm(&stub, &_test_stub, CV_L2) != 0 )
cvtest::cmpEps( Mat(&stub1), Mat(&_test_stub1), &max_diff, 0, &pt, true) < 0 )
{
ts->printf( cvtest::TS::LOG, "readObj method: the read nd matrix is not correct: (%.20g vs %.20g) vs at (%d,%d)\n",
cvGetReal2D(&stub1, pt[0], pt[1]), cvGetReal2D(&_test_stub1, pt[0], pt[1]),
pt[0], pt[1] );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
MatND mat_nd2;
fs["test_mat_nd"] >> mat_nd2;
CvMatND m_nd2 = mat_nd2;
cvGetMat(&m_nd2, &stub, 0, 1);
cvReshape(&stub, &stub1, 1, 0);
if( !CV_ARE_TYPES_EQ(&stub, &_test_stub) ||
!CV_ARE_SIZES_EQ(&stub, &_test_stub) ||
//cvNorm(&stub, &_test_stub, CV_L2) != 0 )
cvtest::cmpEps( Mat(&stub1), Mat(&_test_stub1), &max_diff, 0, &pt, true) < 0 )
{
ts->printf( cvtest::TS::LOG, "C++ method: the read nd matrix is not correct: (%.20g vs %.20g) vs at (%d,%d)\n",
cvGetReal2D(&stub1, pt[0], pt[1]), cvGetReal2D(&_test_stub1, pt[1], pt[0]),
pt[0], pt[1] );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
cvRelease((void**)&m_nd);
Ptr<CvSparseMat> m_s = (CvSparseMat*)fs["test_sparse_mat"].readObj();
Ptr<CvSparseMat> _test_sparse_ = (CvSparseMat*)test_sparse_mat;
Ptr<CvSparseMat> _test_sparse = (CvSparseMat*)cvClone(_test_sparse_);
SparseMat m_s2;
fs["test_sparse_mat"] >> m_s2;
Ptr<CvSparseMat> _m_s2 = (CvSparseMat*)m_s2;
if( !m_s || !CV_IS_SPARSE_MAT(m_s) ||
!cvTsCheckSparse(m_s, _test_sparse,0) ||
!cvTsCheckSparse(_m_s2, _test_sparse,0))
{
ts->printf( cvtest::TS::LOG, "the read sparse matrix is not correct\n" );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
FileNode tl = fs["test_list"];
if( tl.type() != FileNode::SEQ || tl.size() != 6 ||
fabs((double)tl[0] - 0.0000000000001) >= DBL_EPSILON ||
(int)tl[1] != 2 ||
fabs((double)tl[2] - CV_PI) >= DBL_EPSILON ||
(int)tl[3] != -3435345 ||
(string)tl[4] != "2-502 2-029 3egegeg" ||
tl[5].type() != FileNode::MAP || tl[5].size() != 3 ||
(int)tl[5]["month"] != 12 ||
(int)tl[5]["day"] != 31 ||
(int)tl[5]["year"] != 1969 )
{
ts->printf( cvtest::TS::LOG, "the test list is incorrect\n" );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
FileNode tm = fs["test_map"];
FileNode tm_lbp = tm["lbp"];
int real_x = (int)tm["x"];
int real_y = (int)tm["y"];
int real_width = (int)tm["width"];
int real_height = (int)tm["height"];
int real_lbp_val = 0;
FileNodeIterator it;
it = tm_lbp.begin();
real_lbp_val |= (int)*it << 0;
++it;
real_lbp_val |= (int)*it << 1;
it++;
real_lbp_val |= (int)*it << 2;
it += 1;
real_lbp_val |= (int)*it << 3;
FileNodeIterator it2(it);
it2 += 4;
real_lbp_val |= (int)*it2 << 7;
--it2;
real_lbp_val |= (int)*it2 << 6;
it2--;
real_lbp_val |= (int)*it2 << 5;
it2 -= 1;
real_lbp_val |= (int)*it2 << 4;
it2 += -1;
CV_Assert( it == it2 );
if( tm.type() != FileNode::MAP || tm.size() != 5 ||
real_x != 1 ||
real_y != 2 ||
real_width != 100 ||
real_height != 200 ||
tm_lbp.type() != FileNode::SEQ ||
tm_lbp.size() != 8 ||
real_lbp_val != 0xb6 )
{
ts->printf( cvtest::TS::LOG, "the test map is incorrect\n" );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
CvGraph* graph3 = (CvGraph*)fs["test_graph"].readObj();
if(graph2->active_count != vcount || graph3->active_count != vcount ||
graph2->edges->active_count != ecount || graph3->edges->active_count != ecount)
{
ts->printf( cvtest::TS::LOG, "the cloned or read graph have wrong number of vertices or edges\n" );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
for( i = 0; i < ecount; i++ )
{
CvGraphEdge* edge2 = cvFindGraphEdge(graph2, edges[i][0], edges[i][1]);
CvGraphEdge* edge3 = cvFindGraphEdge(graph3, edges[i][0], edges[i][1]);
if( !edge2 || edge2->weight != (float)(i+1) ||
!edge3 || edge3->weight != (float)(i+1) )
{
ts->printf( cvtest::TS::LOG, "the cloned or read graph do not have the edge (%d, %d)\n", edges[i][0], edges[i][1] );
ts->set_failed_test_info( cvtest::TS::FAIL_INVALID_OUTPUT );
return;
}
}
fs.release();
if( !mem )
remove(filename.c_str());
}
}
bool CascadeClassifier::Data::read(const FileNode &root)
{
static const float THRESHOLD_EPS = 1e-5f;
// load stage params
String stageTypeStr = (String)root[CC_STAGE_TYPE];
if( stageTypeStr == CC_BOOST )
stageType = BOOST;
else
return false;
String featureTypeStr = (String)root[CC_FEATURE_TYPE];
if( featureTypeStr == CC_HAAR )
featureType = FeatureEvaluator::HAAR;
else if( featureTypeStr == CC_LBP )
featureType = FeatureEvaluator::LBP;
else if( featureTypeStr == CC_HOG )
featureType = FeatureEvaluator::HOG;
else
return false;
origWinSize.width = (int)root[CC_WIDTH];
origWinSize.height = (int)root[CC_HEIGHT];
CV_Assert( origWinSize.height > 0 && origWinSize.width > 0 );
isStumpBased = (int)(root[CC_STAGE_PARAMS][CC_MAX_DEPTH]) == 1 ? true : false;
// load feature params
FileNode fn = root[CC_FEATURE_PARAMS];
if( fn.empty() )
return false;
ncategories = fn[CC_MAX_CAT_COUNT];
int subsetSize = (ncategories + 31)/32,
nodeStep = 3 + ( ncategories>0 ? subsetSize : 1 );
// load stages
fn = root[CC_STAGES];
if( fn.empty() )
return false;
stages.reserve(fn.size());
classifiers.clear();
nodes.clear();
FileNodeIterator it = fn.begin(), it_end = fn.end();
for( int si = 0; it != it_end; si++, ++it )
{
FileNode fns = *it;
Stage stage;
stage.threshold = (float)fns[CC_STAGE_THRESHOLD] - THRESHOLD_EPS;
fns = fns[CC_WEAK_CLASSIFIERS];
if(fns.empty())
return false;
stage.ntrees = (int)fns.size();
stage.first = (int)classifiers.size();
stages.push_back(stage);
classifiers.reserve(stages[si].first + stages[si].ntrees);
FileNodeIterator it1 = fns.begin(), it1_end = fns.end();
for( ; it1 != it1_end; ++it1 ) // weak trees
{
FileNode fnw = *it1;
FileNode internalNodes = fnw[CC_INTERNAL_NODES];
FileNode leafValues = fnw[CC_LEAF_VALUES];
if( internalNodes.empty() || leafValues.empty() )
return false;
DTree tree;
tree.nodeCount = (int)internalNodes.size()/nodeStep;
classifiers.push_back(tree);
nodes.reserve(nodes.size() + tree.nodeCount);
leaves.reserve(leaves.size() + leafValues.size());
if( subsetSize > 0 )
subsets.reserve(subsets.size() + tree.nodeCount*subsetSize);
FileNodeIterator internalNodesIter = internalNodes.begin(), internalNodesEnd = internalNodes.end();
for( ; internalNodesIter != internalNodesEnd; ) // nodes
{
DTreeNode node;
node.left = (int)*internalNodesIter; ++internalNodesIter;
node.right = (int)*internalNodesIter; ++internalNodesIter;
node.featureIdx = (int)*internalNodesIter; ++internalNodesIter;
if( subsetSize > 0 )
{
for( int j = 0; j < subsetSize; j++, ++internalNodesIter )
subsets.push_back((int)*internalNodesIter);
node.threshold = 0.f;
}
else
{
node.threshold = (float)*internalNodesIter; ++internalNodesIter;
}
nodes.push_back(node);
}
internalNodesIter = leafValues.begin(), internalNodesEnd = leafValues.end();
for( ; internalNodesIter != internalNodesEnd; ++internalNodesIter ) // leaves
leaves.push_back((float)*internalNodesIter);
}
}
return true;
}
int main( int argc, const char** argv )
{
CommandLineParser parser(argc, argv,
"{ help h usage ? | | show this message }"
"{ image i | | (required) path to reference image }"
"{ model m | | (required) path to cascade xml file }"
"{ data d | | (optional) path to video output folder }"
);
// Read in the input arguments
if (parser.has("help")){
parser.printMessage();
printLimits();
return 0;
}
string model(parser.get<string>("model"));
string output_folder(parser.get<string>("data"));
string image_ref = (parser.get<string>("image"));
if (model.empty() || image_ref.empty()){
parser.printMessage();
printLimits();
return -1;
}
// Value for timing
// You can increase this to have a better visualisation during the generation
int timing = 1;
// Value for cols of storing elements
int cols_prefered = 5;
// Open the XML model
FileStorage fs;
bool model_ok = fs.open(model, FileStorage::READ);
if (!model_ok){
cerr << "the cascade file '" << model << "' could not be loaded." << endl;
return -1;
}
// Get a the required information
// First decide which feature type we are using
FileNode cascade = fs["cascade"];
string feature_type = cascade["featureType"];
bool haar = false, lbp = false;
if (feature_type.compare("HAAR") == 0){
haar = true;
}
if (feature_type.compare("LBP") == 0){
lbp = true;
}
if ( feature_type.compare("HAAR") != 0 && feature_type.compare("LBP")){
cerr << "The model is not an HAAR or LBP feature based model!" << endl;
cerr << "Please select a model that can be visualized by the software." << endl;
return -1;
}
// We make a visualisation mask - which increases the window to make it at least a bit more visible
int resize_factor = 10;
int resize_storage_factor = 10;
Mat reference_image = imread(image_ref, IMREAD_GRAYSCALE );
if (reference_image.empty()){
cerr << "the reference image '" << image_ref << "'' could not be loaded." << endl;
return -1;
}
Mat visualization;
resize(reference_image, visualization, Size(reference_image.cols * resize_factor, reference_image.rows * resize_factor));
// First recover for each stage the number of weak features and their index
// Important since it is NOT sequential when using LBP features
vector< vector<int> > stage_features;
FileNode stages = cascade["stages"];
FileNodeIterator it_stages = stages.begin(), it_stages_end = stages.end();
int idx = 0;
for( ; it_stages != it_stages_end; it_stages++, idx++ ){
vector<int> current_feature_indexes;
FileNode weak_classifiers = (*it_stages)["weakClassifiers"];
FileNodeIterator it_weak = weak_classifiers.begin(), it_weak_end = weak_classifiers.end();
vector<int> values;
for(int idy = 0; it_weak != it_weak_end; it_weak++, idy++ ){
(*it_weak)["internalNodes"] >> values;
current_feature_indexes.push_back( (int)values[2] );
}
stage_features.push_back(current_feature_indexes);
}
// If the output option has been chosen than we will store a combined image plane for
// each stage, containing all weak classifiers for that stage.
bool draw_planes = false;
stringstream output_video;
output_video << output_folder << "model_visualization.avi";
VideoWriter result_video;
if( output_folder.compare("") != 0 ){
draw_planes = true;
result_video.open(output_video.str(), VideoWriter::fourcc('X','V','I','D'), 15, Size(reference_image.cols * resize_factor, reference_image.rows * resize_factor), false);
}
if(haar){
// Grab the corresponding features dimensions and weights
FileNode features = cascade["features"];
vector< vector< rect_data > > feature_data;
FileNodeIterator it_features = features.begin(), it_features_end = features.end();
for(int idf = 0; it_features != it_features_end; it_features++, idf++ ){
vector< rect_data > current_feature_rectangles;
FileNode rectangles = (*it_features)["rects"];
int nrects = (int)rectangles.size();
for(int k = 0; k < nrects; k++){
rect_data current_data;
FileNode single_rect = rectangles[k];
current_data.x = (int)single_rect[0];
current_data.y = (int)single_rect[1];
current_data.w = (int)single_rect[2];
current_data.h = (int)single_rect[3];
current_data.weight = (float)single_rect[4];
current_feature_rectangles.push_back(current_data);
}
feature_data.push_back(current_feature_rectangles);
}
// Loop over each possible feature on its index, visualise on the mask and wait a bit,
// then continue to the next feature.
// If visualisations should be stored then do the in between calculations
Mat image_plane;
Mat metadata = Mat::zeros(150, 1000, CV_8UC1);
vector< rect_data > current_rects;
for(int sid = 0; sid < (int)stage_features.size(); sid ++){
if(draw_planes){
int features_nmbr = (int)stage_features[sid].size();
int cols = cols_prefered;
int rows = features_nmbr / cols;
if( (features_nmbr % cols) > 0){
rows++;
}
image_plane = Mat::zeros(reference_image.rows * resize_storage_factor * rows, reference_image.cols * resize_storage_factor * cols, CV_8UC1);
}
for(int fid = 0; fid < (int)stage_features[sid].size(); fid++){
stringstream meta1, meta2;
meta1 << "Stage " << sid << " / Feature " << fid;
meta2 << "Rectangles: ";
Mat temp_window = visualization.clone();
Mat temp_metadata = metadata.clone();
int current_feature_index = stage_features[sid][fid];
current_rects = feature_data[current_feature_index];
Mat single_feature = reference_image.clone();
resize(single_feature, single_feature, Size(), resize_storage_factor, resize_storage_factor);
for(int i = 0; i < (int)current_rects.size(); i++){
rect_data local = current_rects[i];
if(draw_planes){
if(local.weight >= 0){
rectangle(single_feature, Rect(local.x * resize_storage_factor, local.y * resize_storage_factor, local.w * resize_storage_factor, local.h * resize_storage_factor), Scalar(0), FILLED);
}else{
rectangle(single_feature, Rect(local.x * resize_storage_factor, local.y * resize_storage_factor, local.w * resize_storage_factor, local.h * resize_storage_factor), Scalar(255), FILLED);
}
}
Rect part(local.x * resize_factor, local.y * resize_factor, local.w * resize_factor, local.h * resize_factor);
meta2 << part << " (w " << local.weight << ") ";
if(local.weight >= 0){
rectangle(temp_window, part, Scalar(0), FILLED);
}else{
rectangle(temp_window, part, Scalar(255), FILLED);
}
}
imshow("features", temp_window);
putText(temp_window, meta1.str(), Point(15,15), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(255));
result_video.write(temp_window);
// Copy the feature image if needed
if(draw_planes){
single_feature.copyTo(image_plane(Rect(0 + (fid%cols_prefered)*single_feature.cols, 0 + (fid/cols_prefered) * single_feature.rows, single_feature.cols, single_feature.rows)));
}
putText(temp_metadata, meta1.str(), Point(15,15), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(255));
putText(temp_metadata, meta2.str(), Point(15,40), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(255));
imshow("metadata", temp_metadata);
waitKey(timing);
}
//Store the stage image if needed
if(draw_planes){
stringstream save_location;
save_location << output_folder << "stage_" << sid << ".png";
imwrite(save_location.str(), image_plane);
}
}
}
if(lbp){
// Grab the corresponding features dimensions and weights
FileNode features = cascade["features"];
vector<Rect> feature_data;
FileNodeIterator it_features = features.begin(), it_features_end = features.end();
for(int idf = 0; it_features != it_features_end; it_features++, idf++ ){
FileNode rectangle = (*it_features)["rect"];
Rect current_feature ((int)rectangle[0], (int)rectangle[1], (int)rectangle[2], (int)rectangle[3]);
feature_data.push_back(current_feature);
}
// Loop over each possible feature on its index, visualise on the mask and wait a bit,
// then continue to the next feature.
Mat image_plane;
Mat metadata = Mat::zeros(150, 1000, CV_8UC1);
for(int sid = 0; sid < (int)stage_features.size(); sid ++){
if(draw_planes){
int features_nmbr = (int)stage_features[sid].size();
int cols = cols_prefered;
int rows = features_nmbr / cols;
if( (features_nmbr % cols) > 0){
rows++;
}
image_plane = Mat::zeros(reference_image.rows * resize_storage_factor * rows, reference_image.cols * resize_storage_factor * cols, CV_8UC1);
}
for(int fid = 0; fid < (int)stage_features[sid].size(); fid++){
stringstream meta1, meta2;
meta1 << "Stage " << sid << " / Feature " << fid;
meta2 << "Rectangle: ";
Mat temp_window = visualization.clone();
Mat temp_metadata = metadata.clone();
int current_feature_index = stage_features[sid][fid];
Rect current_rect = feature_data[current_feature_index];
Mat single_feature = reference_image.clone();
resize(single_feature, single_feature, Size(), resize_storage_factor, resize_storage_factor);
// VISUALISATION
// The rectangle is the top left one of a 3x3 block LBP constructor
Rect resized(current_rect.x * resize_factor, current_rect.y * resize_factor, current_rect.width * resize_factor, current_rect.height * resize_factor);
meta2 << resized;
// Top left
rectangle(temp_window, resized, Scalar(255), 1);
// Top middle
rectangle(temp_window, Rect(resized.x + resized.width, resized.y, resized.width, resized.height), Scalar(255), 1);
// Top right
rectangle(temp_window, Rect(resized.x + 2*resized.width, resized.y, resized.width, resized.height), Scalar(255), 1);
// Middle left
rectangle(temp_window, Rect(resized.x, resized.y + resized.height, resized.width, resized.height), Scalar(255), 1);
// Middle middle
rectangle(temp_window, Rect(resized.x + resized.width, resized.y + resized.height, resized.width, resized.height), Scalar(255), FILLED);
// Middle right
rectangle(temp_window, Rect(resized.x + 2*resized.width, resized.y + resized.height, resized.width, resized.height), Scalar(255), 1);
// Bottom left
rectangle(temp_window, Rect(resized.x, resized.y + 2*resized.height, resized.width, resized.height), Scalar(255), 1);
// Bottom middle
rectangle(temp_window, Rect(resized.x + resized.width, resized.y + 2*resized.height, resized.width, resized.height), Scalar(255), 1);
// Bottom right
rectangle(temp_window, Rect(resized.x + 2*resized.width, resized.y + 2*resized.height, resized.width, resized.height), Scalar(255), 1);
if(draw_planes){
Rect resized_inner(current_rect.x * resize_storage_factor, current_rect.y * resize_storage_factor, current_rect.width * resize_storage_factor, current_rect.height * resize_storage_factor);
// Top left
rectangle(single_feature, resized_inner, Scalar(255), 1);
// Top middle
rectangle(single_feature, Rect(resized_inner.x + resized_inner.width, resized_inner.y, resized_inner.width, resized_inner.height), Scalar(255), 1);
// Top right
rectangle(single_feature, Rect(resized_inner.x + 2*resized_inner.width, resized_inner.y, resized_inner.width, resized_inner.height), Scalar(255), 1);
// Middle left
rectangle(single_feature, Rect(resized_inner.x, resized_inner.y + resized_inner.height, resized_inner.width, resized_inner.height), Scalar(255), 1);
// Middle middle
rectangle(single_feature, Rect(resized_inner.x + resized_inner.width, resized_inner.y + resized_inner.height, resized_inner.width, resized_inner.height), Scalar(255), FILLED);
// Middle right
rectangle(single_feature, Rect(resized_inner.x + 2*resized_inner.width, resized_inner.y + resized_inner.height, resized_inner.width, resized_inner.height), Scalar(255), 1);
// Bottom left
rectangle(single_feature, Rect(resized_inner.x, resized_inner.y + 2*resized_inner.height, resized_inner.width, resized_inner.height), Scalar(255), 1);
// Bottom middle
rectangle(single_feature, Rect(resized_inner.x + resized_inner.width, resized_inner.y + 2*resized_inner.height, resized_inner.width, resized_inner.height), Scalar(255), 1);
// Bottom right
rectangle(single_feature, Rect(resized_inner.x + 2*resized_inner.width, resized_inner.y + 2*resized_inner.height, resized_inner.width, resized_inner.height), Scalar(255), 1);
single_feature.copyTo(image_plane(Rect(0 + (fid%cols_prefered)*single_feature.cols, 0 + (fid/cols_prefered) * single_feature.rows, single_feature.cols, single_feature.rows)));
}
putText(temp_metadata, meta1.str(), Point(15,15), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(255));
putText(temp_metadata, meta2.str(), Point(15,40), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(255));
imshow("metadata", temp_metadata);
imshow("features", temp_window);
putText(temp_window, meta1.str(), Point(15,15), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(255));
result_video.write(temp_window);
waitKey(timing);
}
//Store the stage image if needed
if(draw_planes){
stringstream save_location;
save_location << output_folder << "stage_" << sid << ".png";
imwrite(save_location.str(), image_plane);
}
}
}
return 0;
}
