potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-30 15:53:03 +04:00
Code Packages Releases Wiki Activity

Merge remote-tracking branch 'upstream/3.4' into merge-3.4

Browse Source
This commit is contained in:
Alexander Alekhin
2018-07-09 19:31:45 +03:00
parent 43f821afb9 7fe0727930
commit 2da96be217
148 changed files with 2773 additions and 2038 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/ml/src/ann_mlp.cpp
+2 -2
View File
@@ -362,7 +362,7 @@ public:
}
cv::AutoBuffer<double> _buf(buf_sz+noutputs);
double* buf = _buf;
double* buf = _buf.data();
if( !_outputs.needed() )
{
@@ -924,7 +924,7 @@ public:
_idx[i] = i;
AutoBuffer<double> _buf(max_lsize*2);
double* buf[] = { _buf, (double*)_buf + max_lsize };
double* buf[] = { _buf.data(), _buf.data() + max_lsize };
const double* sw = _sw.empty() ? 0 : _sw.ptr<double>();
modules/ml/src/boost.cpp
+1 -1
View File
@@ -205,7 +205,7 @@ public:
int nvars = (int)varIdx.size();
double sumw = 0., C = 1.;
cv::AutoBuffer<double> buf(n + nvars);
double* result = buf;
double* result = buf.data();
float* sbuf = (float*)(result + n);
Mat sample(1, nvars, CV_32F, sbuf);
int predictFlags = bparams.boostType == Boost::DISCRETE ? (PREDICT_MAX_VOTE | RAW_OUTPUT) : PREDICT_SUM;
modules/ml/src/data.cpp
+1 -1
View File
@@ -335,7 +335,7 @@ public:
CatMapHash ofshash;
AutoBuffer<uchar> buf(nsamples);
Mat non_missing(layout == ROW_SAMPLE ? Size(1, nsamples) : Size(nsamples, 1), CV_8U, (uchar*)buf);
Mat non_missing(layout == ROW_SAMPLE ? Size(1, nsamples) : Size(nsamples, 1), CV_8U, buf.data());
bool haveMissing = !missing.empty();
if( haveMissing )
{
modules/ml/src/kdtree.cpp
+3 -3
View File
@@ -285,13 +285,13 @@ int KDTree::findNearest(InputArray _vec, int K, int emax,
CV_Assert(K > 0 && (normType == NORM_L2 || normType == NORM_L1));
AutoBuffer<uchar> _buf((K+1)*(sizeof(float) + sizeof(int)));
int* idx = (int*)(uchar*)_buf;
int* idx = (int*)_buf.data();
float* dist = (float*)(idx + K + 1);
int i, j, ncount = 0, e = 0;
int qsize = 0, maxqsize = 1 << 10;
AutoBuffer<uchar> _pqueue(maxqsize*sizeof(PQueueElem));
PQueueElem* pqueue = (PQueueElem*)(uchar*)_pqueue;
PQueueElem* pqueue = (PQueueElem*)_pqueue.data();
emax = std::max(emax, 1);
for( e = 0; e < emax; )
@@ -433,7 +433,7 @@ void KDTree::findOrthoRange(InputArray _lowerBound,
std::vector<int> idx;
AutoBuffer<int> _stack(MAX_TREE_DEPTH*2 + 1);
int* stack = _stack;
int* stack = _stack.data();
int top = 0;
stack[top++] = 0;
modules/ml/src/knearest.cpp
+1 -1
View File
@@ -149,7 +149,7 @@ public:
int k = std::min(k0, nsamples);
AutoBuffer<float> buf(testcount*k*2);
float* dbuf = buf;
float* dbuf = buf.data();
float* rbuf = dbuf + testcount*k;
const float* rptr = responses.ptr<float>();
modules/ml/src/nbayes.cpp
+2 -2
View File
@@ -241,8 +241,8 @@ public:
}
// allocate memory and initializing headers for calculating
cv::AutoBuffer<double> _buffer(nvars*2);
double* _diffin = _buffer;
double* _diffout = _buffer + nvars;
double* _diffin = _buffer.data();
double* _diffout = _buffer.data() + nvars;
Mat diffin( 1, nvars, CV_64FC1, _diffin );
Mat diffout( 1, nvars, CV_64FC1, _diffout );
modules/ml/src/svm.cpp
+2 -2
View File
@@ -1579,7 +1579,7 @@ public:
return;
AutoBuffer<double> vbuf(var_count);
double* v = vbuf;
double* v = vbuf.data();
Mat new_sv(df_count, var_count, CV_32F);
vector<DecisionFunc> new_df;
@@ -1914,7 +1914,7 @@ public:
int class_count = !svm->class_labels.empty() ? (int)svm->class_labels.total() : svmType == ONE_CLASS ? 1 : 0;
AutoBuffer<float> _buffer(sv_total + (class_count+1)*2);
float* buffer = _buffer;
float* buffer = _buffer.data();
int i, j, dfi, k, si;
modules/ml/src/tree.cpp
+12 -12
View File
@@ -417,7 +417,7 @@ int DTreesImpl::findBestSplit( const vector<int>& _sidx )
int splitidx = -1;
int vi_, nv = (int)activeVars.size();
AutoBuffer<int> buf(w->maxSubsetSize*2);
int *subset = buf, *best_subset = subset + w->maxSubsetSize;
int *subset = buf.data(), *best_subset = subset + w->maxSubsetSize;
WSplit split, best_split;
best_split.quality = 0.;
@@ -488,7 +488,7 @@ void DTreesImpl::calcValue( int nidx, const vector<int>& _sidx )
// misclassified samples with cv_labels(*)==j.
// compute the number of instances of each class
double* cls_count = buf;
double* cls_count = buf.data();
double* cv_cls_count = cls_count + m;
double max_val = -1, total_weight = 0;
@@ -592,7 +592,7 @@ void DTreesImpl::calcValue( int nidx, const vector<int>& _sidx )
}
else
{
double *cv_sum = buf, *cv_sum2 = cv_sum + cv_n;
double *cv_sum = buf.data(), *cv_sum2 = cv_sum + cv_n;
double* cv_count = (double*)(cv_sum2 + cv_n);
for( j = 0; j < cv_n; j++ )
@@ -646,7 +646,7 @@ DTreesImpl::WSplit DTreesImpl::findSplitOrdClass( int vi, const vector<int>& _si
const int* sidx = &_sidx[0];
const int* responses = &w->cat_responses[0];
const double* weights = &w->sample_weights[0];
double* lcw = (double*)(uchar*)buf;
double* lcw = (double*)buf.data();
double* rcw = lcw + m;
float* values = (float*)(rcw + m);
int* sorted_idx = (int*)(values + n);
@@ -717,7 +717,7 @@ void DTreesImpl::clusterCategories( const double* vectors, int n, int m, double*
int iters = 0, max_iters = 100;
int i, j, idx;
cv::AutoBuffer<double> buf(n + k);
double *v_weights = buf, *c_weights = buf + n;
double *v_weights = buf.data(), *c_weights = buf.data() + n;
bool modified = true;
RNG r((uint64)-1);
@@ -819,12 +819,12 @@ DTreesImpl::WSplit DTreesImpl::findSplitCatClass( int vi, const vector<int>& _si
base_size += mi;
AutoBuffer<double> buf(base_size + n);
double* lc = (double*)buf;
double* lc = buf.data();
double* rc = lc + m;
double* _cjk = rc + m*2, *cjk = _cjk;
double* c_weights = cjk + m*mi;
int* labels = (int*)(buf + base_size);
int* labels = (int*)(buf.data() + base_size);
w->data->getNormCatValues(vi, _sidx, labels);
const int* responses = &w->cat_responses[0];
const double* weights = &w->sample_weights[0];
@@ -991,7 +991,7 @@ DTreesImpl::WSplit DTreesImpl::findSplitOrdReg( int vi, const vector<int>& _sidx
AutoBuffer<uchar> buf(n*(sizeof(int) + sizeof(float)));
float* values = (float*)(uchar*)buf;
float* values = (float*)buf.data();
int* sorted_idx = (int*)(values + n);
w->data->getValues(vi, _sidx, values);
const double* responses = &w->ord_responses[0];
@@ -1053,7 +1053,7 @@ DTreesImpl::WSplit DTreesImpl::findSplitCatReg( int vi, const vector<int>& _sidx
int mi = getCatCount(vi);
AutoBuffer<double> buf(3*mi + 3 + n);
double* sum = (double*)buf + 1;
double* sum = buf.data() + 1;
double* counts = sum + mi + 1;
double** sum_ptr = (double**)(counts + mi);
int* cat_labels = (int*)(sum_ptr + mi);
@@ -1148,7 +1148,7 @@ int DTreesImpl::calcDir( int splitidx, const vector<int>& _sidx,
if( mi <= 0 ) // split on an ordered variable
{
float c = split.c;
float* values = buf;
float* values = buf.data();
w->data->getValues(vi, _sidx, values);
for( i = 0; i < n; i++ )
@@ -1169,7 +1169,7 @@ int DTreesImpl::calcDir( int splitidx, const vector<int>& _sidx,
else
{
const int* subset = &w->wsubsets[split.subsetOfs];
int* cat_labels = (int*)(float*)buf;
int* cat_labels = (int*)buf.data();
w->data->getNormCatValues(vi, _sidx, cat_labels);
for( i = 0; i < n; i++ )
@@ -1372,7 +1372,7 @@ float DTreesImpl::predictTrees( const Range& range, const Mat& sample, int flags
int i, ncats = (int)catOfs.size(), nclasses = (int)classLabels.size();
int catbufsize = ncats > 0 ? nvars : 0;
AutoBuffer<int> buf(nclasses + catbufsize + 1);
int* votes = buf;
int* votes = buf.data();
int* catbuf = votes + nclasses;
const int* cvidx = (flags & (COMPRESSED_INPUT|PREPROCESSED_INPUT)) == 0 && !varIdx.empty() ? &compVarIdx[0] : 0;
const uchar* vtype = &varType[0];