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

- Choosing better CE-CLM models for C++ (more general and work well in a broader setting)

Browse Source 
- Reducing the memory load of the model in C++ by half
This commit is contained in:
Tadas Baltrusaitis
2017-09-13 19:39:11 +01:00
parent 99a5fa160b
commit f5b364c54b
37 changed files with 278 additions and 108 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
lib/local/LandmarkDetector/src/Patch_experts.cpp
View File

@@ -226,9 +226,15 @@ void Patch_experts::Response(vector<cv::Mat_<float> >& patch_expert_responses, c
// Get intensity response either from the SVR, CCNF, or CEN patch experts (prefer CEN as they are the most accurate so far)
if (!cen_expert_intensity.empty())
{
cen_expert_intensity[scale][view_id][i].ResponseSparse(area_of_interest, patch_expert_responses[i], interp_mat);
// For space and memory saving use a mirrored patch expert
if(!cen_expert_intensity[scale][view_id][i].biases.empty())
{
cen_expert_intensity[scale][view_id][i].ResponseSparse(area_of_interest, patch_expert_responses[i], interp_mat);
}
else
{
cen_expert_intensity[scale][mirror_views.at<int>(view_id)][mirror_inds.at<int>(i)].ResponseSparse_mirror(area_of_interest, patch_expert_responses[i], interp_mat);
}
// A slower, but slightly more accurate version
//cen_expert_intensity[scale][view_id][i].Response(area_of_interest, patch_expert_responses[i]);
@@ -546,6 +552,9 @@ void Patch_experts::Read_CEN_patch_experts(string expert_location, std::vector<c
}
int numberOfPoints = visibility[0].rows;
LandmarkDetector::ReadMatBin(patchesFile, mirror_inds);
LandmarkDetector::ReadMatBin(patchesFile, mirror_views);
// read the patches themselves
for (size_t i = 0; i < patches.size(); i++)
{