This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
hi. You can use opencv_traincascade instead of opencv_haartraining. opencv_traincascade supports both Haar & LBP features.LPB is several time faster than Haar. In addition opencv_traincascade can use TBB in multi core system. In the first you have to sample down original image. You can set four parameters in haar cascade in OpenCV that affects the speed 1- Scale Factor: how much the image size is reduced at each image scale (The speed-accuracy trade-off) 2- Min Neighbor 3- min Size (minimum size of face) 4 -max Size (minimum size of face)
| | 2 | No.2 Revision |
hi.
You can use opencv_traincascade instead of opencv_haartraining. opencv_traincascade supports both Haar & LBP features.LPB is several time faster than Haar. In addition opencv_traincascade can use TBB in multi core system.
In the first you have to down sample down original image.
You can set four parameters in haar cascade in OpenCV that affects the speed
1- Scale Factor: how much the image size is reduced at each image
scale (The speed-accuracy trade-off)
2- Min Neighbor
3- min Size (minimum size of face)
4 -max Size (minimum size of face)
