教学文库网 - 权威文档分享云平台
您的当前位置:首页 > 文库大全 > 专业资料 >

基于皮肤模板和改进HMM的自动人脸识别系统(4)

来源:网络收集 时间:2026-08-04
导读: formanceof improvedHMMiscomparablewiththatofDCTHMM whenthenumber oftrainingimagesis 5. 基于皮肤模板和改进HMM的自动人

formanceof

improvedHMMiscomparablewiththatofDCTHMM

whenthenumber

oftrainingimagesis

5.

基于皮肤模板和改进HMM的自动人脸识别系统

74

深圳大学学报理工版

第25卷

Table1

ExperimentalresultsforORL

database表1ORL数据库的实验结果Inthe

next

experiment,acolorface

imagedatabaseof320

imagesisproducedto

test

theeffectoftheskinmaskingproce-

dure.AcreativeWebCamwasusedand10framepictures

are

captured

automaticallyforeachpersonwhentheysitbefore

the

camera.Thevolunteerscan

tilt

orrotatetheirheadwithin±100.

Theface

detection

meduledescribedinsection2isused

to

detect

thefacesautomaticallywithdifferentlighting,posesandexpres—

sions.The

volunteersinvolved

are

fromdifferentcountriesand

thuscapturedfaceimages

cover

different

typesofskincolors.Fig

6showstheflowcharttogeneratetheNottinghamfacedatabases.DatabasesNott_AandNott—B

ore

produced

atthesametimedur-

ingthecapturingprocess.Comparedwith

databaseNott.—A,the

skinmaskingmoduledescribedin

section2wasusedforcreating

databaseNoRB.APentium-41.8

GHzPC

was噼edinexperi—

mentandtheiiTlagesizeis320×240.111efacedetectionprocesstookabout100m8andtheskinmasking

processtookabout45

Ins.

Fig.6

Theflowcharttogeneratefacedatabase

图6人脸数据库生成流程

BothHMMfacerecognitionmethodsaretestedusingdata-baseNotLAandNottB.t11eresultsaresummarizedintable2.

Inthis

experiment,5ofthefaceimagesarerandomlychosen

to

traintheHMMmodelforeachpersonandtheremaining5faceimages

ale

usedfortesting.Otherparametersarekeptthesame

asthoseinthepreviousexperimentsusingORLdatabase.From

table2,one

can

observethatbothalgorithmsachievebetter

per-

formancewhendatabaseNott—Bisusedfor

testing.About5%

improvementsareachievedfortheimprovedHMM.Sincedata baseNott—Bis

just

skin

maskedversionofdatabaseNott—A,it

call

beconcludedthattheskinmaskingmodule

can

effectively

Table2

RecognitionRatefordatabase

Nott_AandNoR—B

表2数据库Nott—A和NoR—B的实验结果

万 

方数据improvetheperformanceofthefacerecognitionsystem.Bycom—

parison

amongthese

recognitionalgorithms,onealsoobservethat

thealgorithmofimprovedHMMachievesmuchbetterperformance

thanothersfor

bothdatabaseNott_AandNott_B.

Conclusion

Afullyworkingautomatedfacerecognitionsystem

hasbeen

implementedinthepaperFaceimagesaredetectedfromthelive

videostreamandidentified

against

developed

database.Afast

skincolorbasedfacemaskingmethodhasbeenproposedto

im-

provethedetection

accuracy.Thepaper

hasalsoproposed

DWT

basedfacerecognitionHMM,whichimprovedtheaccuracyof

o-

riginalDCTbasedHMM.

The

facedetectionmodule

can

processup

to10frames

per

secondfromvideostreams.However,sincerecognitionrequiresmuchmorecomputationthan

detection,asimple

facetracking

moduleisintegratedintothesystem.Foreachvideoframe,faces

are

detectedand

comparedwiththerecognizedfacesintheprevi—

OUS

frame.Ifthe

distanceo …… 此处隐藏:1649字,全部文档内容请下载后查看。喜欢就下载吧 ……

基于皮肤模板和改进HMM的自动人脸识别系统(4).doc 将本文的Word文档下载到电脑,方便复制、编辑、收藏和打印
本文链接:https://www.jiaowen.net/wenku/268230.html(转载请注明文章来源)
Copyright © 2020-2025 教文网 版权所有
声明 :本网站尊重并保护知识产权,根据《信息网络传播权保护条例》,如果我们转载的作品侵犯了您的权利,请在一个月内通知我们,我们会及时删除。
客服QQ:78024566 邮箱:78024566@qq.com
苏ICP备19068818号-2
Top
× 游客快捷下载通道(下载后可以自由复制和排版)
VIP包月下载
特价:29 元/月 原价:99元
低至 0.3 元/份 每月下载150
全站内容免费自由复制
VIP包月下载
特价:29 元/月 原价:99元
低至 0.3 元/份 每月下载150
全站内容免费自由复制
注:下载文档有可能出现无法下载或内容有问题,请联系客服协助您处理。
× 常见问题(客服时间:周一到周五 9:30-18:00)