tailieunhanh - O’Reilly Learning OpenCV phần 10

CvDTreeParams:: use_surrogates, mà sẽ đảm bảo rằng các đặc tính thay thế chia được lưu trữ tại mỗi nút. Một tùy chọn quan trọng là sử dụng priors để thiết lập các "chi phí" dương tính giả. Một lần nữa, nếu chúng ta đang học nấm ăn được hoặc độc hại sau đó chúng ta có thể thiết lập priors priors float [] = {1,0, 10,0} | that maybe missing then we can set use_surrogates to CvDTreeParams use_surrogates which will ensure that alternate features on which the splitting is based are stored at each node. An important option is that of using priors to set the cost of false positives. Again if we are learning edible or poisonous mushrooms then we might set the priors to be float priorsf then each error of labeling a poisonous mushroom ediblewould cost ten timesas much aslabeling an ediblemusheoom poisonous. The CoBoost etass codieiwslhe member sealewh ichie aCt See eointur to the weak slase sifiore thst inheriSs Sesne CsaTree decieion sa LovscBoost avid GenSlsSsott the treso ase rasrsss iow teoes shousSCat coatect fleetsvd-aaiet volues ldeciscan trees fc r ether mtnhods reSurnonly votes for dase s idp-sio-vo Of sSaes 1 so negetive . TVcsoan-laineOaloss t eoue nee hrs stew follewiwvpuotatype class so c seCTsee . i public CslssltTrcet ie sidont scvBoosOereeO vữttsl bse tcsine s dis -OriiUnhaOa Cinst sobsasdt-idx CvBoost ensemble Cl. lishsnl VOID secush double s iisCsnlvolD tecus O zdeStsrage. fs CiFdeMusId node CtBect01 essence ClBTsecTra i n D a ta -dirte visCsnl vtid ItooteoCid OoOlnsTe ettooblot to Training is almost the same as for decision trees but there is an extra parameter called dssceCcedn of weok olassifieoe feem ecratch. If update is set to true 1 then we just add new weak classifiers onto the existing group. The function prototype for training a boosted classifier is Note test fol cemputer sisis. festuros are omoeted from as imas- andtàes fe- totlis dessi r sence cdny ese c1masinevet missi ng Uiisise fsaeu resarice cots ie dets ioHectsd eyhamans esr exampin furnetting ie tskt ths J en nn Semse cturn idd Icy. t l o lln oOSse tloKcts is KePtTh oseieeoitivs. tie oiscs ootype OeBeoes if oho bsoseed tree d-ssi-filer. The objects of type CvBoostTree are the weak classifiers that constitute the overall boosted strong classifier. .