[*] Fixed #974 ("GPU CascadeClassifier fails with some training files"): Moved IsNodeLeaf bit from NodeDescriptor to FeatureDescriptor for both left and right nodes, therefore from now on max number of rects in a feature is 31

pull/13383/head
Anton Obukhov 14 years ago
parent 2388fa223e
commit 58476b64a6
  1. 70
      modules/gpu/src/cascadeclassifier.cpp
  2. 44
      modules/gpu/src/nvidia/NCVHaarObjectDetection.cu
  3. 39
      modules/gpu/src/nvidia/NCVHaarObjectDetection.hpp

@ -336,24 +336,24 @@ NCVStatus loadFromXML(const std::string &filename,
haar.NumClassifierTotalNodes = 0;
haar.NumFeatures = 0;
haar.ClassifierSize.width = 0;
haar.ClassifierSize.height = 0;
haar.ClassifierSize.height = 0;
haar.bHasStumpsOnly = true;
haar.bNeedsTiltedII = false;
Ncv32u curMaxTreeDepth;
std::vector<char> xmlFileCont;
std::vector<char> xmlFileCont;
std::vector<HaarClassifierNode128> h_TmpClassifierNotRootNodes;
haarStages.resize(0);
haarClassifierNodes.resize(0);
haarFeatures.resize(0);
haarFeatures.resize(0);
Ptr<CvHaarClassifierCascade> oldCascade = (CvHaarClassifierCascade*)cvLoad(filename.c_str(), 0, 0, 0);
if (oldCascade.empty())
return NCV_HAAR_XML_LOADING_EXCEPTION;
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
haar.ClassifierSize.width = oldCascade->orig_window_size.width;
haar.ClassifierSize.height = oldCascade->orig_window_size.height;
@ -366,53 +366,58 @@ NCVStatus loadFromXML(const std::string &filename,
curStage.setStageThreshold(oldCascade->stage_classifier[s].threshold);
int treesCount = oldCascade->stage_classifier[s].count;
for(int t = 0; t < treesCount; ++t) // bytrees
{
for(int t = 0; t < treesCount; ++t) // by trees
{
Ncv32u nodeId = 0;
CvHaarClassifier* tree = &oldCascade->stage_classifier[s].classifier[t];
int nodesCount = tree->count;
for(int n = 0; n < nodesCount; ++n) //by features
{
for(int n = 0; n < nodesCount; ++n) //by features
{
CvHaarFeature* feature = &tree->haar_feature[n];
HaarClassifierNode128 curNode;
HaarClassifierNode128 curNode;
curNode.setThreshold(tree->threshold[n]);
NcvBool bIsLeftNodeLeaf = false;
NcvBool bIsRightNodeLeaf = false;
HaarClassifierNodeDescriptor32 nodeLeft;
if ( tree->left[n] <= 0 )
{
Ncv32f leftVal = tree->alpha[-tree->left[n]];
ncvStat = nodeLeft.create(leftVal);
ncvAssertReturn(ncvStat == NCV_SUCCESS, ncvStat);
ncvAssertReturn(ncvStat == NCV_SUCCESS, ncvStat);
bIsLeftNodeLeaf = true;
}
else
{
Ncv32u leftNodeOffset = tree->left[n];
Ncv32u leftNodeOffset = tree->left[n];
nodeLeft.create((Ncv32u)(h_TmpClassifierNotRootNodes.size() + leftNodeOffset - 1));
haar.bHasStumpsOnly = false;
}
curNode.setLeftNodeDesc(nodeLeft);
HaarClassifierNodeDescriptor32 nodeRight;
if ( tree->right[n] <= 0 )
{
Ncv32f rightVal = tree->alpha[-tree->right[n]];
{
Ncv32f rightVal = tree->alpha[-tree->right[n]];
ncvStat = nodeRight.create(rightVal);
ncvAssertReturn(ncvStat == NCV_SUCCESS, ncvStat);
bIsRightNodeLeaf = true;
}
else
{
Ncv32u rightNodeOffset = tree->right[n];
{
Ncv32u rightNodeOffset = tree->right[n];
nodeRight.create((Ncv32u)(h_TmpClassifierNotRootNodes.size() + rightNodeOffset - 1));
haar.bHasStumpsOnly = false;
}
curNode.setRightNodeDesc(nodeRight);
curNode.setRightNodeDesc(nodeRight);
Ncv32u tiltedVal = feature->tilted;
haar.bNeedsTiltedII = (tiltedVal != 0);
haar.bNeedsTiltedII = (tiltedVal != 0);
Ncv32u featureId = 0;
Ncv32u featureId = 0;
for(int l = 0; l < CV_HAAR_FEATURE_MAX; ++l) //by rects
{
Ncv32u rectX = feature->rect[l].r.x;
@ -435,7 +440,8 @@ NCVStatus loadFromXML(const std::string &filename,
}
HaarFeatureDescriptor32 tmpFeatureDesc;
ncvStat = tmpFeatureDesc.create(haar.bNeedsTiltedII, featureId, haarFeatures.size() - featureId);
ncvStat = tmpFeatureDesc.create(haar.bNeedsTiltedII, bIsLeftNodeLeaf, bIsRightNodeLeaf,
featureId, haarFeatures.size() - featureId);
ncvAssertReturn(NCV_SUCCESS == ncvStat, ncvStat);
curNode.setFeatureDesc(tmpFeatureDesc);
@ -453,14 +459,14 @@ NCVStatus loadFromXML(const std::string &filename,
}
nodeId++;
}
}
}
curStage.setNumClassifierRootNodes(treesCount);
haarStages.push_back(curStage);
haarStages.push_back(curStage);
}
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
//fill in cascade stats
haar.NumStages = haarStages.size();
@ -472,8 +478,10 @@ NCVStatus loadFromXML(const std::string &filename,
Ncv32u offsetRoot = haarClassifierNodes.size();
for (Ncv32u i=0; i<haarClassifierNodes.size(); i++)
{
HaarFeatureDescriptor32 featureDesc = haarClassifierNodes[i].getFeatureDesc();
HaarClassifierNodeDescriptor32 nodeLeft = haarClassifierNodes[i].getLeftNodeDesc();
if (!nodeLeft.isLeaf())
if (!featureDesc.isLeftNodeLeaf())
{
Ncv32u newOffset = nodeLeft.getNextNodeOffset() + offsetRoot;
nodeLeft.create(newOffset);
@ -481,7 +489,7 @@ NCVStatus loadFromXML(const std::string &filename,
haarClassifierNodes[i].setLeftNodeDesc(nodeLeft);
HaarClassifierNodeDescriptor32 nodeRight = haarClassifierNodes[i].getRightNodeDesc();
if (!nodeRight.isLeaf())
if (!featureDesc.isRightNodeLeaf())
{
Ncv32u newOffset = nodeRight.getNextNodeOffset() + offsetRoot;
nodeRight.create(newOffset);
@ -490,8 +498,10 @@ NCVStatus loadFromXML(const std::string &filename,
}
for (Ncv32u i=0; i<h_TmpClassifierNotRootNodes.size(); i++)
{
HaarFeatureDescriptor32 featureDesc = h_TmpClassifierNotRootNodes[i].getFeatureDesc();
HaarClassifierNodeDescriptor32 nodeLeft = h_TmpClassifierNotRootNodes[i].getLeftNodeDesc();
if (!nodeLeft.isLeaf())
if (!featureDesc.isLeftNodeLeaf())
{
Ncv32u newOffset = nodeLeft.getNextNodeOffset() + offsetRoot;
nodeLeft.create(newOffset);
@ -499,7 +509,7 @@ NCVStatus loadFromXML(const std::string &filename,
h_TmpClassifierNotRootNodes[i].setLeftNodeDesc(nodeLeft);
HaarClassifierNodeDescriptor32 nodeRight = h_TmpClassifierNotRootNodes[i].getRightNodeDesc();
if (!nodeRight.isLeaf())
if (!featureDesc.isRightNodeLeaf())
{
Ncv32u newOffset = nodeRight.getNextNodeOffset() + offsetRoot;
nodeRight.create(newOffset);

@ -444,10 +444,22 @@ __global__ void applyHaarClassifierAnchorParallel(Ncv32u *d_IImg, Ncv32u IImgStr
HaarClassifierNodeDescriptor32 nodeLeft = curNode.getLeftNodeDesc();
HaarClassifierNodeDescriptor32 nodeRight = curNode.getRightNodeDesc();
Ncv32f nodeThreshold = curNode.getThreshold();
HaarClassifierNodeDescriptor32 nextNodeDescriptor;
nextNodeDescriptor = (curNodeVal < scaleArea * pixelStdDev * nodeThreshold) ? nodeLeft : nodeRight;
NcvBool nextNodeIsLeaf;
if (curNodeVal < scaleArea * pixelStdDev * nodeThreshold)
{
nextNodeDescriptor = nodeLeft;
nextNodeIsLeaf = featuresDesc.isLeftNodeLeaf();
}
else
{
nextNodeDescriptor = nodeRight;
nextNodeIsLeaf = featuresDesc.isRightNodeLeaf();
}
if (nextNodeDescriptor.isLeaf())
if (nextNodeIsLeaf)
{
Ncv32f tmpLeafValue = nextNodeDescriptor.getLeafValue();
curStageSum += tmpLeafValue;
@ -572,10 +584,22 @@ __global__ void applyHaarClassifierClassifierParallel(Ncv32u *d_IImg, Ncv32u IIm
HaarClassifierNodeDescriptor32 nodeLeft = curNode.getLeftNodeDesc();
HaarClassifierNodeDescriptor32 nodeRight = curNode.getRightNodeDesc();
Ncv32f nodeThreshold = curNode.getThreshold();
HaarClassifierNodeDescriptor32 nextNodeDescriptor;
nextNodeDescriptor = (curNodeVal < scaleArea * pixelStdDev * nodeThreshold) ? nodeLeft : nodeRight;
NcvBool nextNodeIsLeaf;
if (nextNodeDescriptor.isLeaf())
if (curNodeVal < scaleArea * pixelStdDev * nodeThreshold)
{
nextNodeDescriptor = nodeLeft;
nextNodeIsLeaf = featuresDesc.isLeftNodeLeaf();
}
else
{
nextNodeDescriptor = nodeRight;
nextNodeIsLeaf = featuresDesc.isRightNodeLeaf();
}
if (nextNodeIsLeaf)
{
Ncv32f tmpLeafValue = nextNodeDescriptor.getLeafValue();
curStageSum += tmpLeafValue;
@ -2135,8 +2159,9 @@ NCVStatus ncvApplyHaarClassifierCascade_host(NCVMatrix<Ncv32u> &h_integralImage,
while (bMoreNodesToTraverse)
{
HaarClassifierNode128 curNode = h_HaarNodes.ptr()[curNodeOffset];
Ncv32u curNodeFeaturesNum = curNode.getFeatureDesc().getNumFeatures();
Ncv32u curNodeFeaturesOffs = curNode.getFeatureDesc().getFeaturesOffset();
HaarFeatureDescriptor32 curFeatDesc = curNode.getFeatureDesc();
Ncv32u curNodeFeaturesNum = curFeatDesc.getNumFeatures();
Ncv32u curNodeFeaturesOffs = curFeatDesc.getFeaturesOffset();
Ncv32f curNodeVal = 0.f;
for (Ncv32u iRect=0; iRect<curNodeFeaturesNum; iRect++)
@ -2161,19 +2186,22 @@ NCVStatus ncvApplyHaarClassifierCascade_host(NCVMatrix<Ncv32u> &h_integralImage,
HaarClassifierNodeDescriptor32 nodeLeft = curNode.getLeftNodeDesc();
HaarClassifierNodeDescriptor32 nodeRight = curNode.getRightNodeDesc();
Ncv32f nodeThreshold = curNode.getThreshold();
HaarClassifierNodeDescriptor32 nextNodeDescriptor;
NcvBool nextNodeIsLeaf;
if (curNodeVal < scaleAreaPixels * h_weights.ptr()[i * h_weights.stride() + j] * nodeThreshold)
{
nextNodeDescriptor = nodeLeft;
nextNodeIsLeaf = curFeatDesc.isLeftNodeLeaf();
}
else
{
nextNodeDescriptor = nodeRight;
nextNodeIsLeaf = curFeatDesc.isRightNodeLeaf();
}
NcvBool tmpIsLeaf = nextNodeDescriptor.isLeaf();
if (tmpIsLeaf)
if (nextNodeIsLeaf)
{
Ncv32f tmpLeafValue = nextNodeDescriptor.getLeafValueHost();
curStageSum += tmpLeafValue;

@ -112,7 +112,9 @@ struct HaarFeatureDescriptor32
private:
#define HaarFeatureDescriptor32_Interpret_MaskFlagTilted 0x80000000
#define HaarFeatureDescriptor32_CreateCheck_MaxNumFeatures 0x7F
#define HaarFeatureDescriptor32_Interpret_MaskFlagLeftNodeLeaf 0x40000000
#define HaarFeatureDescriptor32_Interpret_MaskFlagRightNodeLeaf 0x20000000
#define HaarFeatureDescriptor32_CreateCheck_MaxNumFeatures 0x1F
#define HaarFeatureDescriptor32_NumFeatures_Shift 24
#define HaarFeatureDescriptor32_CreateCheck_MaxFeatureOffset 0x00FFFFFF
@ -120,7 +122,8 @@ private:
public:
__host__ NCVStatus create(NcvBool bTilted, Ncv32u numFeatures, Ncv32u offsetFeatures)
__host__ NCVStatus create(NcvBool bTilted, NcvBool bLeftLeaf, NcvBool bRightLeaf,
Ncv32u numFeatures, Ncv32u offsetFeatures)
{
if (numFeatures > HaarFeatureDescriptor32_CreateCheck_MaxNumFeatures)
{
@ -132,6 +135,8 @@ public:
}
this->desc = 0;
this->desc |= (bTilted ? HaarFeatureDescriptor32_Interpret_MaskFlagTilted : 0);
this->desc |= (bLeftLeaf ? HaarFeatureDescriptor32_Interpret_MaskFlagLeftNodeLeaf : 0);
this->desc |= (bRightLeaf ? HaarFeatureDescriptor32_Interpret_MaskFlagRightNodeLeaf : 0);
this->desc |= (numFeatures << HaarFeatureDescriptor32_NumFeatures_Shift);
this->desc |= offsetFeatures;
return NCV_SUCCESS;
@ -142,9 +147,19 @@ public:
return (this->desc & HaarFeatureDescriptor32_Interpret_MaskFlagTilted) != 0;
}
__device__ __host__ NcvBool isLeftNodeLeaf(void)
{
return (this->desc & HaarFeatureDescriptor32_Interpret_MaskFlagLeftNodeLeaf) != 0;
}
__device__ __host__ NcvBool isRightNodeLeaf(void)
{
return (this->desc & HaarFeatureDescriptor32_Interpret_MaskFlagRightNodeLeaf) != 0;
}
__device__ __host__ Ncv32u getNumFeatures(void)
{
return (this->desc & ~HaarFeatureDescriptor32_Interpret_MaskFlagTilted) >> HaarFeatureDescriptor32_NumFeatures_Shift;
return (this->desc >> HaarFeatureDescriptor32_NumFeatures_Shift) & HaarFeatureDescriptor32_CreateCheck_MaxNumFeatures;
}
__device__ __host__ Ncv32u getFeaturesOffset(void)
@ -158,34 +173,18 @@ struct HaarClassifierNodeDescriptor32
{
uint1 _ui1;
#define HaarClassifierNodeDescriptor32_Interpret_MaskSwitch (1 << 30)
__host__ NCVStatus create(Ncv32f leafValue)
{
if ((*(Ncv32u *)&leafValue) & HaarClassifierNodeDescriptor32_Interpret_MaskSwitch)
{
return NCV_HAAR_XML_LOADING_EXCEPTION;
}
*(Ncv32f *)&this->_ui1 = leafValue;
return NCV_SUCCESS;
}
__host__ NCVStatus create(Ncv32u offsetHaarClassifierNode)
{
if (offsetHaarClassifierNode >= HaarClassifierNodeDescriptor32_Interpret_MaskSwitch)
{
return NCV_HAAR_XML_LOADING_EXCEPTION;
}
this->_ui1.x = offsetHaarClassifierNode;
this->_ui1.x |= HaarClassifierNodeDescriptor32_Interpret_MaskSwitch;
return NCV_SUCCESS;
}
__device__ __host__ NcvBool isLeaf(void)
{
return !(this->_ui1.x & HaarClassifierNodeDescriptor32_Interpret_MaskSwitch);
}
__host__ Ncv32f getLeafValueHost(void)
{
return *(Ncv32f *)&this->_ui1.x;
@ -200,7 +199,7 @@ struct HaarClassifierNodeDescriptor32
__device__ __host__ Ncv32u getNextNodeOffset(void)
{
return (this->_ui1.x & ~HaarClassifierNodeDescriptor32_Interpret_MaskSwitch);
return this->_ui1.x;
}
};

Loading…
Cancel
Save