The programs feature extraction from positive and negative sample images of the HOG and LBP, samples using support vector machine training, be for pedestrian classification. Detection using the trained classifier, experimental results show that the method can effectively detect pedestrians in the image, and to achieve better results. Has passed the tests.




File list:
HOG-LBP detection
................\compareAll.m
................\Demotest.m
................\Demotrain.m
................\hogcalculator.m
................\hogfeat.m
................\HOGGradient.m
................\HOGSVM1.m
................\HOG_LBP.m
................\HOG_LBPSVM1.m
................\LBP.m
................\lbp_com.m
................\location.m
................\main.m
................\make.m
................\max_per5.m
................\realdetection.m
................\rectangleonimage.m
................\SVM2.m
................\test.m
................\winslide.m