ZF/OpenFace
1
0
Fork 0
mirror of https://gitcode.com/gh_mirrors/ope/OpenFace.git synced 2026-05-16 04:08:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

Parallelizing the MTCNN face detection

Browse Source
This commit is contained in:
Tadas Baltrusaitis
2017-08-17 14:17:20 +01:00
parent 441777c9ed
commit a3be476d4d
1 changed files with 59 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

77
lib/local/LandmarkDetector/src/FaceDetectorMTCNN.cpp
View File

@@ -348,7 +348,6 @@ void convolution_single_kern_fft(const vector<cv::Mat_<float> >& input_imgs, vec
dft(dst, dst, 0, _templs[k].rows);
}
// TODO is a deep copy needed
_templ_dfts[dftsize.width] = dftTempl;
}
@@ -546,7 +545,6 @@ void convolution_fft(std::vector<cv::Mat_<float> >& outputs, const std::vector<c
cv::Mat integral_image;
cv::Mat integral_image_sq;
// TODO can TBB-ify this
for (size_t k = 0; k < kernels[in].size(); ++k)
{
cv::Mat_<float> kernel = kernels[in][k];
@@ -1098,7 +1096,13 @@ bool FaceDetectorMTCNN::DetectFaces(vector<cv::Rect_<double> >& o_regions, const
vector<float> scores_all;
vector<cv::Rect_<float> > proposal_corrections_all;
for (int i = 0; i < num_scales; ++i)
// As the scales will be done in parallel have some containers for them
vector<vector<cv::Rect_<float> > > proposal_boxes_cross_scale(num_scales);
vector<vector<float> > scores_cross_scale(num_scales);
vector<vector<cv::Rect_<float> > > proposal_corrections_cross_scale(num_scales);
tbb::parallel_for(0, (int)num_scales, [&](int i) {
//for (int i = 0; i < num_scales; ++i)
{
double scale = ((double)face_support / (double)min_face_size)*cv::pow(pyramid_factor, i);
@@ -1114,7 +1118,7 @@ bool FaceDetectorMTCNN::DetectFaces(vector<cv::Rect_<double> >& o_regions, const
// Actual PNet CNN step
std::vector<cv::Mat_<float> > pnet_out = PNet.Inference(normalised_img, true);
// Clear the precomputations, as the image sizes will be different (TODO could be useful for videos)
// Clear the precomputations, as the image sizes will be different
PNet.ClearPrecomp();
// Extract the probabilities from PNet response
@@ -1131,14 +1135,21 @@ bool FaceDetectorMTCNN::DetectFaces(vector<cv::Rect_<double> >& o_regions, const
vector<cv::Rect_<float> > proposal_corrections;
generate_bounding_boxes(proposal_boxes, scores, proposal_corrections, prob_heatmap, corrections_heatmap, scale, t1, face_support);
// Perform non-maximum supression on proposals in this scale
vector<int> to_keep = non_maximum_supression(proposal_boxes, scores, 0.5, false);
select_subset(to_keep, proposal_boxes, scores, proposal_corrections);
proposal_boxes_cross_scale[i] = proposal_boxes;
scores_cross_scale[i] = scores;
proposal_corrections_cross_scale[i] = proposal_corrections;
}
});
proposal_boxes_all.insert(proposal_boxes_all.end(), proposal_boxes.begin(), proposal_boxes.end());
scores_all.insert(scores_all.end(), scores.begin(), scores.end());
proposal_corrections_all.insert(proposal_corrections_all.end(), proposal_corrections.begin(), proposal_corrections.end());
// Perform non-maximum supression on proposals accross scales and combine them
for (int i = 0; i < num_scales; ++i)
{
vector<int> to_keep = non_maximum_supression(proposal_boxes_cross_scale[i], scores_cross_scale[i], 0.5, false);
select_subset(to_keep, proposal_boxes_cross_scale[i], scores_cross_scale[i], proposal_corrections_cross_scale[i]);
proposal_boxes_all.insert(proposal_boxes_all.end(), proposal_boxes_cross_scale[i].begin(), proposal_boxes_cross_scale[i].end());
scores_all.insert(scores_all.end(), scores_cross_scale[i].begin(), scores_cross_scale[i].end());
proposal_corrections_all.insert(proposal_corrections_all.end(), proposal_corrections_cross_scale[i].begin(), proposal_corrections_cross_scale[i].end());
}
// Preparation for RNet step
@@ -1153,8 +1164,9 @@ bool FaceDetectorMTCNN::DetectFaces(vector<cv::Rect_<double> >& o_regions, const
rectify(proposal_boxes_all);
// Creating proposal images from previous step detections
to_keep.clear();
for (size_t k = 0; k < proposal_boxes_all.size(); ++k)
vector<bool> above_thresh(proposal_boxes_all.size());
tbb::parallel_for(0, (int)proposal_boxes_all.size(), [&](int k) {
//for (size_t k = 0; k < proposal_boxes_all.size(); ++k)
{
float width_target = proposal_boxes_all[k].width + 1;
float height_target = proposal_boxes_all[k].height + 1;
@@ -1192,10 +1204,24 @@ bool FaceDetectorMTCNN::DetectFaces(vector<cv::Rect_<double> >& o_regions, const
proposal_corrections_all[k].height = rnet_out[0].at<float>(5);
if(prob >= t2)
{
to_keep.push_back(k);
above_thresh[k] = true;
}
else
{
above_thresh[k] = false;
}
}
});
to_keep.clear();
for (size_t i = 0; i < above_thresh.size(); ++i)
{
if (above_thresh[i])
{
to_keep.push_back(i);
}
}
// Pick only the bounding boxes above the threshold
select_subset(to_keep, proposal_boxes_all, scores_all, proposal_corrections_all);
@@ -1210,9 +1236,10 @@ bool FaceDetectorMTCNN::DetectFaces(vector<cv::Rect_<double> >& o_regions, const
rectify(proposal_boxes_all);
// Preparing for the ONet stage
to_keep.clear();
for (size_t k = 0; k < proposal_boxes_all.size(); ++k)
above_thresh.clear();
above_thresh.resize(proposal_boxes_all.size());
tbb::parallel_for(0, (int)proposal_boxes_all.size(), [&](int k) {
//for (size_t k = 0; k < proposal_boxes_all.size(); ++k)
{
float width_target = proposal_boxes_all[k].width + 1;
float height_target = proposal_boxes_all[k].height + 1;
@@ -1250,7 +1277,21 @@ bool FaceDetectorMTCNN::DetectFaces(vector<cv::Rect_<double> >& o_regions, const
proposal_corrections_all[k].height = onet_out[0].at<float>(5);
if (prob >= t3)
{
to_keep.push_back(k);
above_thresh[k] = true;
}
else
{
above_thresh[k] = false;
}
}
});
to_keep.clear();
for (size_t i = 0; i < above_thresh.size(); ++i)
{
if (above_thresh[i])
{
to_keep.push_back(i);
}
}