【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总

实验环境

=====================================================================================

system: ubuntu 16.04

cpu: 8  Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz

QMake version 3.0  Qt version 5.6.2

opencv 3.3.0

cmake version 3.5.1

Python 3.6.1 |Anaconda custom (64-bit)

dlib  19.10.0

image size 640*480

=====================================================================================

实验结果

(一)左:cmake-dlib / 右:cmake-opencv

 【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总

(二)qmake-dlib(左:debug;右:release)

【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总

(三)qmake-opencv(左:debug;右:release)

【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总

python-dlib

【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总

(四)python-opencv

【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总

=====================================================================================

列表总结640*480 opcv-detector和dlib-detector Time情况

【dlib opencv - detector landmark】 ubuntu上针对dlib-hog和opencv haar人脸检测与landmar-68在不同平台上运行时间实验结果汇总

Time all dlib-128d-5point                                                                                                                 11-18ms

Time all dlib-128d-68point                                                                                                               11-20ms