Tomoaki Teshima
63f5f93063
fix test failure on ODROID-N2
5 years ago
mshabunin
3e28d51779
Fixed several OpenCL compiler warnings
8 years ago
Vadim Pisarevsky
5a94a95fbf
improvements in Haar CascadeClassifier: 1) use CV_32S instead of CV_32F for the integral of squares (which is more accurate and more efficient); 2) skip the window if its contrast is too low
10 years ago
Yan Wang
433c3da7cf
Optimize the performance of cascade OpenCL kernel.
...
1. Use built-in mad() instead of += and *.
2. For stump stages, if weight.z == 0, avoid unnecessary calculation
because some features only have 2 rectangles..
It could improve
OCL_Cascade_Image_MinSize_CascadeClassifier.CascadeClassifier/*
about 10% mean values.
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
10 years ago
Yan Wang
99aaf4030a
Remove unnecessary barrier in runHaarClassifier().
...
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
10 years ago
Yan Wang
df697f6520
Optimize runLBPClassifierStumpSimple by built-in mad24.
...
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
10 years ago
Yan Wang
f5b21e8141
Avoid deadlock becasue some work item couldn't reach barrier in loop
...
and unsynchronized.
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
10 years ago
Yan Wang
07ab0c1c53
Avoid compile unnecessary OpenCL kernel.
...
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
10 years ago
Yan Wang
9699fde8d2
Use hypot to do vector calculation.
...
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
10 years ago
Yan Wang
efa84d8225
Use preprocessor for constant values in OpenCL kernel instead of
...
the parameter variable.
It could improve the performance of
OCL_Cascade_Image_MinSize_CascadeClassifier.CascadeClassifier/*.
Especially,
OCL_Cascade_Image_MinSize_CascadeClassifier.CascadeClassifier/15
OCL_Cascade_Image_MinSize_CascadeClassifier.CascadeClassifier/16
could be improved about 2% in Intel platform.
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
10 years ago
Alexander Karsakov
b14c314fc3
Fixed incorrect thread synchronizations
11 years ago
Alexander Karsakov
30a8308f8e
Enabled Intel-specific optimizations for HOG detector.
11 years ago
Alexander Karsakov
4698d62125
Removed workaround for Intel platform.
11 years ago
Vadim Pisarevsky
de4b1c668d
hopefully fixed compile warnings and the doc builder warnings
11 years ago
Vadim Pisarevsky
30593ee55e
3rd attempt to prepare patch with improved OpenCL kernels of CascadeClassifier.
11 years ago
Konstantin Matskevich
9a62df1650
HOG
11 years ago
Vadim Pisarevsky
8998186ce4
removed extra whitespaces and hopefully fixed the test failures
11 years ago
Vadim Pisarevsky
d084d19779
added OpenCL optimization for LBP-based face detector
11 years ago
Vadim Pisarevsky
1540910542
started adding OpenCL acceleration of LBP-based object detectors
11 years ago
Vadim Pisarevsky
f41f633d2d
fixed various warnings from the “doc” and other builders
11 years ago
Vadim Pisarevsky
ec3f22cee2
renamed OpenCL kernel filename; made some final changes
11 years ago
Vadim Pisarevsky
b7553d4e2e
some attempts to tune the performance
11 years ago
Vadim Pisarevsky
02fb3f0a77
it finally works!!!
11 years ago
Vadim Pisarevsky
302a5adcc2
converted Haar cascades to the new format; now they are handled with C++ code.
11 years ago
Vadim Pisarevsky
8762ee3f6f
removed unnecessary opencl kernels
11 years ago
Vadim Pisarevsky
d914f20a4c
updated patch to bring in the first functions with "transparent API"
11 years ago