• lucky.lu
    了解作者
  • Visual C++
    开发工具
  • 7.2MB
    文件大小
  • rar
    文件格式
  • 0
    收藏次数
  • 1 积分
    下载积分
  • 6
    下载次数
  • 2018-03-27 17:34
    上传日期
人脸图像识别,很不错的程序,移植一下就可以用在项目上
Fatigue-detection.rar
  • 4
  • Debug
  • vc60.pdb
    404KB
  • vc60.idb
    329KB
  • camerads.opt
    42.5KB
  • camerads.ilk
    685KB
  • camerads.pdb
    865KB
  • CameraDS.obj
    112.1KB
  • main.obj
    221.5KB
  • camerads.exe
    524.1KB
  • camerads.pch
    7.5MB
  • DirectShow
  • Include
  • d3d.h
    86KB
  • mpconfig.h
    6.2KB
  • d3dx8math.h
    37.9KB
  • dmodshow.h
    4.3KB
  • Mpeg2Bits.h
    1.5KB
  • Mpeg2Error.h
    4KB
  • aviriff.h
    13.5KB
  • tuner.h
    395.9KB
  • d3dx9mesh.h
    77.1KB
  • il21dec.h
    4KB
  • dshowasf.h
    9.4KB
  • d3dx.h
    575B
  • d3dxshapes.h
    9.9KB
  • amaudio.h
    1.8KB
  • mediaobj.h
    45.5KB
  • control.h
    174.1KB
  • Segment.h
    551.9KB
  • dv.h
    2.2KB
  • Amvideo.h
    19.4KB
  • dmerror.h
    26.7KB
  • dmdls.h
    7.4KB
  • d3dx9effect.h
    36.1KB
  • d3d8caps.h
    15.7KB
  • d3drmwin.h
    1.1KB
  • ksuuids.h
    8.7KB
  • dsound.h
    107.5KB
  • qedit.h
    302.2KB
  • dmoreg.h
    4KB
  • Mstve.h
    420.6KB
  • atsmedia.h
    1.1KB
  • ksmedia.h
    217.4KB
  • dvp.h
    32.2KB
  • d3dx9shader.h
    35.2KB
  • comlite.h
    944B
  • dsconf.h
    9.6KB
  • dxfile.h
    8KB
  • ddraw.h
    239.5KB
  • d3d8types.h
    62.2KB
  • d3dx8effect.h
    7.1KB
  • ddstream.h
    16.8KB
  • Msvidctl.tlb
    105.5KB
  • ksproxy.h
    21.9KB
  • ks.h
    145.3KB
  • d3dx8tex.h
    54.7KB
  • d3dx8.h
    893B
  • d3d9caps.h
    21.6KB
  • dmusics.h
    9.4KB
  • d3dxmath.inl
    36.3KB
  • DxDiag.h
    8KB
  • evcode.h
    17.5KB
  • strmif.h
    888.5KB
  • mpeg2data.h
    29.9KB
  • bdamedia.h
    53.3KB
  • videoacc.h
    22KB
  • d3dx8math.inl
    35.4KB
  • d3dxsprite.h
    15.6KB
  • dplay8.h
    69KB
  • d3dvec.inl
    5.3KB
  • dpnathlp.h
    13.4KB
  • d3dx8shape.h
    7.8KB
  • amva.h
    3.9KB
  • uuids.h
    63.7KB
  • vpnotify.h
    2.4KB
  • vpconfig.h
    4.8KB
  • dvdevcod.h
    15.7KB
  • mediaerr.h
    640B
  • dmusicc.h
    35.4KB
  • qnetwork.h
    17.4KB
  • amparse.h
    1KB
  • d3dxerr.h
    7.7KB
  • ksguid.h
    637B
  • dmksctrl.h
    5.2KB
  • d3dx9tex.h
    59.7KB
  • Msvidctl.h
    72.2KB
  • dls2.h
    4.7KB
  • errors.h
    1.2KB
  • bdatypes.h
    15.1KB
  • xprtdefs.h
    26.6KB
  • d3dxcore.h
    45.4KB
  • dxtrans.h
    155.2KB
  • dls1.h
    9.1KB
  • austream.h
    20.1KB
  • tune.h
    40.8KB
  • edevdefs.h
    26.3KB
  • edevctrl.h
    10.6KB
  • d3dx8mesh.h
    28.4KB
  • dxva.h
    32.5KB
内容介绍
<html xmlns="http://www.w3.org/1999/xhtml"> <head> <meta charset="utf-8"> <meta name="generator" content="pdf2htmlEX"> <meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1"> <link rel="stylesheet" href="https://static.pudn.com/base/css/base.min.css"> <link rel="stylesheet" href="https://static.pudn.com/base/css/fancy.min.css"> <link rel="stylesheet" href="https://static.pudn.com/prod/directory_preview_static/62758a6b16f2c0769c05e0b7/raw.css"> <script src="https://static.pudn.com/base/js/compatibility.min.js"></script> <script src="https://static.pudn.com/base/js/pdf2htmlEX.min.js"></script> <script> try{ pdf2htmlEX.defaultViewer = new pdf2htmlEX.Viewer({}); }catch(e){} </script> <title></title> </head> <body> <div id="sidebar" style="display: none"> <div id="outline"> </div> </div> <div id="pf1" class="pf w0 h0" data-page-no="1"><div class="pc pc1 w0 h0"><img class="bi x0 y0 w1 h1" alt="" src="https://static.pudn.com/prod/directory_preview_static/62758a6b16f2c0769c05e0b7/bg1.jpg"><div class="t m0 x1 h2 y1 ff1 fs0 fc0 sc0 ls0 ws0">A<span class="_ _0"></span>verage<span class="_ _1"> </span>of<span class="_ _1"> </span>Synthetic<span class="_ _2"> </span>Exact<span class="_ _1"> </span>Filters</div><div class="t m0 x2 h3 y2 ff2 fs1 fc0 sc0 ls0 ws0">David<span class="_ _3"> </span>S.<span class="_ _3"> </span>Bolme<span class="_ _4"> </span>Bruce<span class="_ _3"> </span>A.<span class="_ _1"> </span>Draper<span class="_ _5"> </span>J.<span class="_ _3"> </span>Ross<span class="_ _1"> </span>Be<span class="_ _6"></span>veridge</div><div class="t m0 x3 h3 y3 ff2 fs1 fc0 sc0 ls0 ws0">Colorado<span class="_ _3"> </span>State<span class="_ _1"> </span>Uni<span class="_ _6"></span>versity</div><div class="t m0 x4 h3 y4 ff2 fs1 fc0 sc0 ls0 ws0">Fort<span class="_ _3"> </span>Collins,<span class="_ _7"> </span>CO<span class="_ _3"> </span>80521,<span class="_ _7"> </span>USA</div><div class="t m0 x5 h4 y5 ff3 fs2 fc0 sc0 ls0 ws0">bolme@cs.colostate.edu</div><div class="t m0 x6 h5 y6 ff1 fs1 fc0 sc0 ls0 ws0">Abstract</div><div class="t m0 x7 h6 y7 ff4 fs3 fc0 sc0 ls0 ws0">This<span class="_ _8"> </span>paper<span class="_ _8"> </span>intr<span class="_ _6"></span>oduces<span class="_ _8"> </span>a<span class="_ _8"> </span>class<span class="_ _8"> </span>of<span class="_ _8"> </span>correlation<span class="_ _8"> </span>&#64257;lter<span class="_ _6"></span>s<span class="_ _8"> </span>called</div><div class="t m0 x8 h6 y8 ff4 fs3 fc0 sc0 ls0 ws0">A<span class="_ _6"></span>verag<span class="_ _6"></span>e<span class="_ _2"> </span>of<span class="_ _1"> </span>Synthetic<span class="_ _1"> </span>Exact<span class="_ _1"> </span>F<span class="_ _9"></span>ilters<span class="_ _1"> </span>(ASEF).<span class="_ _1"> </span>F<span class="_ _a"></span>or<span class="_ _1"> </span>ASEF<span class="_ _0"></span>,<span class="_ _1"> </span>the</div><div class="t m0 x8 h6 y9 ff4 fs3 fc0 sc0 ls0 ws0">corr<span class="_ _6"></span>elation<span class="_ _3"> </span>output<span class="_ _3"> </span>is<span class="_ _3"> </span>completely<span class="_ _3"> </span>speci&#64257;ed<span class="_ _3"> </span>for<span class="_ _3"> </span>each<span class="_ _3"> </span>training</div><div class="t m0 x8 h6 ya ff4 fs3 fc0 sc0 ls0 ws0">image<span class="_ _6"></span>.<span class="_ _2"> </span>This<span class="_ _3"> </span>is<span class="_ _3"> </span>in<span class="_ _3"> </span>mark<span class="_ _6"></span>ed<span class="_ _3"> </span>contrast<span class="_ _3"> </span>to<span class="_ _3"> </span>prior<span class="_ _8"> </span>methods<span class="_ _3"> </span>such<span class="_ _3"> </span>as</div><div class="t m0 x8 h6 yb ff4 fs3 fc0 sc0 ls0 ws0">Synthetic<span class="_ _3"> </span>Discriminant<span class="_ _3"> </span>Functions<span class="_ _3"> </span>(SDFs)<span class="_ _3"> </span>which<span class="_ _3"> </span>only<span class="_ _3"> </span>spec-</div><div class="t m0 x8 h6 yc ff4 fs3 fc0 sc0 ls0 ws0">ify<span class="_ _3"> </span>a<span class="_ _3"> </span>single<span class="_ _8"> </span>output<span class="_ _3"> </span>value<span class="_ _3"> </span>per<span class="_ _3"> </span>tr<span class="_ _6"></span>aining<span class="_ _3"> </span>image<span class="_ _9"></span>.<span class="_ _1"> </span>Advantages<span class="_ _3"> </span>of</div><div class="t m0 x8 h6 yd ff4 fs3 fc0 sc0 ls0 ws0">ASEF<span class="_ _1"> </span>tr<span class="_ _9"></span>aining<span class="_ _1"> </span>include:<span class="_ _b"> </span>insenitivity<span class="_ _1"> </span>to<span class="_ _7"> </span>over-&#64257;tting,<span class="_ _1"> </span>gr<span class="_ _6"></span>eater</div><div class="t m0 x8 h6 ye ff4 fs3 fc0 sc0 ls0 ws0">&#64258;exibility<span class="_ _7"> </span>with<span class="_ _7"> </span>re<span class="_ _6"></span>gar<span class="_ _6"></span>d<span class="_ _1"> </span>to<span class="_ _7"> </span>training<span class="_ _7"> </span>images,<span class="_ _1"> </span>and<span class="_ _7"> </span>more<span class="_ _7"> </span>r<span class="_ _6"></span>obust</div><div class="t m0 x8 h6 yf ff4 fs3 fc0 sc0 ls0 ws0">behavior<span class="_ _2"> </span>in<span class="_ _2"> </span>the<span class="_ _2"> </span>pr<span class="_ _6"></span>esence<span class="_ _c"> </span>of<span class="_ _2"> </span>structur<span class="_ _6"></span>ed<span class="_ _2"> </span>backgr<span class="_ _6"></span>ounds.<span class="_ _d"> </span>The</div><div class="t m0 x8 h6 y10 ff4 fs3 fc0 sc0 ls0 ws0">theory<span class="_ _c"> </span>and<span class="_ _c"> </span>design<span class="_ _e"> </span>of<span class="_ _c"> </span>ASEF<span class="_ _c"> </span>&#64257;lters<span class="_ _c"> </span>is<span class="_ _e"> </span>pr<span class="_ _6"></span>esented<span class="_ _e"> </span>using<span class="_ _c"> </span>eye</div><div class="t m0 x8 h6 y11 ff4 fs3 fc0 sc0 ls0 ws0">localization<span class="_ _2"> </span>on<span class="_ _c"> </span>the<span class="_ _2"> </span>FERET<span class="_ _c"> </span>database<span class="_ _2"> </span>as<span class="_ _c"> </span>an<span class="_ _2"> </span>example<span class="_ _2"> </span>task.</div><div class="t m0 x8 h6 y12 ff4 fs3 fc0 sc0 ls0 ws0">ASEF<span class="_ _1"> </span>is<span class="_ _1"> </span>compared<span class="_ _1"> </span>to<span class="_ _1"> </span>other<span class="_ _1"> </span>popular<span class="_ _1"> </span>correlation<span class="_ _1"> </span>&#64257;lters<span class="_ _1"> </span>in-</div><div class="t m0 x8 h6 y13 ff4 fs3 fc0 sc0 ls0 ws0">cluding<span class="_ _8"> </span>SDF<span class="_ _0"></span>,<span class="_ _8"> </span>MACE,<span class="_ _8"> </span>O<span class="_ _6"></span>TF<span class="_ _0"></span>,<span class="_ _8"> </span>and<span class="_ _8"> </span>UMACE,<span class="_ _8"> </span>and<span class="_ _8"> </span>with<span class="_ _8"> </span>other<span class="_ _8"> </span>e<span class="_ _6"></span>ye</div><div class="t m0 x8 h6 y14 ff4 fs3 fc0 sc0 ls0 ws0">localization<span class="_ _8"> </span>methods<span class="_ _8"> </span>including<span class="_ _8"> </span>Gabor<span class="_ _8"> </span>J<span class="_ _6"></span>ets<span class="_ _8"> </span>and<span class="_ _8"> </span>the<span class="_ _8"> </span>OpenCV</div><div class="t m0 x8 h6 y15 ff4 fs3 fc0 sc0 ls0 ws0">Cascade<span class="_ _3"> </span>Classi&#64257;er<span class="_ _a"></span>.<span class="_ _2"> </span>ASEF<span class="_ _3"> </span>is<span class="_ _3"> </span>shown<span class="_ _3"> </span>to<span class="_ _7"> </span>outperform<span class="_ _3"> </span>all<span class="_ _3"> </span>these</div><div class="t m0 x8 h6 y16 ff4 fs3 fc0 sc0 ls0 ws0">methods,<span class="_ _8"> </span>locating<span class="_ _8"> </span>the<span class="_ _3"> </span>e<span class="_ _6"></span>ye<span class="_ _8"> </span>to<span class="_ _3"> </span>within<span class="_ _8"> </span>the<span class="_ _8"> </span>radius<span class="_ _8"> </span>of<span class="_ _8"> </span>the<span class="_ _8"> </span>iris<span class="_ _8"> </span>ap-</div><div class="t m0 x8 h6 y17 ff4 fs3 fc0 sc0 ls0 ws0">pr<span class="_ _6"></span>oximately<span class="_ _3"> </span>98.5%<span class="_ _8"> </span>of<span class="_ _3"> </span>the<span class="_ _8"> </span>time.</div><div class="t m0 x8 h5 y18 ff1 fs1 fc0 sc0 ls0 ws0">1.<span class="_ _3"> </span>Introduction</div><div class="t m0 x7 h7 y19 ff2 fs3 fc0 sc0 ls0 ws0">A<span class="_ _7"> </span>common<span class="_ _1"> </span>way<span class="_ _7"> </span>to<span class="_ _7"> </span>detect<span class="_ _1"> </span>patterns<span class="_ _7"> </span>in<span class="_ _1"> </span>images<span class="_ _7"> </span>is<span class="_ _1"> </span>through</div><div class="t m0 x8 h7 y1a ff2 fs3 fc0 sc0 ls0 ws0">correlation<span class="_ _1"> </span>with<span class="_ _2"> </span>an<span class="_ _2"> </span>example<span class="_ _1"> </span>template<span class="_ _2"> </span>[4].<span class="_ _d"> </span>The<span class="_ _1"> </span>simplicity</div><div class="t m0 x8 h7 y1b ff2 fs3 fc0 sc0 ls0 ws0">and<span class="_ _8"> </span>ef<span class="_ _6"></span>&#64257;ciency<span class="_ _f"> </span>of<span class="_ _f"> </span>this<span class="_ _f"> </span>approach<span class="_ _f"> </span>continually<span class="_ _8"> </span>draws<span class="_ _f"> </span>researcher</div><div class="t m0 x8 h7 y1c ff2 fs3 fc0 sc0 ls0 ws0">attention<span class="_ _e"> </span>despite<span class="_ _e"> </span>many<span class="_ _e"> </span>known<span class="_ _e"> </span>weaknesses,<span class="_ _10"> </span>and<span class="_ _e"> </span>a<span class="_ _e"> </span>steady</div><div class="t m0 x8 h7 y1d ff2 fs3 fc0 sc0 ls0 ws0">progression<span class="_ _2"> </span>of<span class="_ _2"> </span>advances<span class="_ _2"> </span>has<span class="_ _2"> </span>continued<span class="_ _2"> </span>to<span class="_ _2"> </span>breath<span class="_ _c"> </span>new<span class="_ _1"> </span>life</div><div class="t m0 x8 h7 y1e ff2 fs3 fc0 sc0 ls0 ws0">into<span class="_ _8"> </span>this<span class="_ _3"> </span>well<span class="_ _8"> </span>established<span class="_ _3"> </span>technique.</div><div class="t m0 x7 h7 y1f ff2 fs3 fc0 sc0 ls0 ws0">One<span class="_ _8"> </span>commonly<span class="_ _8"> </span>recognized<span class="_ _8"> </span>weakness<span class="_ _3"> </span>of<span class="_ _f"> </span>simple<span class="_ _8"> </span>template</div><div class="t m0 x8 h7 y20 ff2 fs3 fc0 sc0 ls0 ws0">matching<span class="_ _2"> </span>is<span class="_ _c"> </span>that<span class="_ _2"> </span>while<span class="_ _c"> </span>the<span class="_ _c"> </span>response<span class="_ _2"> </span>to<span class="_ _c"> </span>a<span class="_ _c"> </span>perfect<span class="_ _2"> </span>example</div><div class="t m0 x8 h7 y21 ff2 fs3 fc0 sc0 ls0 ws0">of<span class="_ _e"> </span>the<span class="_ _e"> </span>template<span class="_ _b"> </span>pattern<span class="_ _e"> </span>will<span class="_ _e"> </span>always<span class="_ _e"> </span>be<span class="_ _e"> </span>high,<span class="_ _10"> </span>the<span class="_ _e"> </span>relativ<span class="_ _6"></span>e</div><div class="t m0 x8 h7 y22 ff2 fs3 fc0 sc0 ls0 ws0">strength<span class="_ _1"> </span>of<span class="_ _7"> </span>responses<span class="_ _1"> </span>to<span class="_ _1"> </span>alternativ<span class="_ _6"></span>e<span class="_ _1"> </span>patterns<span class="_ _1"> </span>can<span class="_ _1"> </span>be<span class="_ _7"> </span>unpre-</div><div class="t m0 x8 h7 y23 ff2 fs3 fc0 sc0 ls0 ws0">dictable.<span class="_ _1"> </span>A<span class="_ _3"> </span>family<span class="_ _3"> </span>of<span class="_ _3"> </span>correlation<span class="_ _8"> </span>&#64257;lters<span class="_ _3"> </span>has<span class="_ _3"> </span>been<span class="_ _3"> </span>de<span class="_ _6"></span>veloped</div><div class="t m0 x8 h7 y24 ff2 fs3 fc0 sc0 ls0 ws0">that<span class="_ _1"> </span>endeav<span class="_ _9"></span>ors<span class="_ _1"> </span>to<span class="_ _1"> </span>overcome<span class="_ _1"> </span>this<span class="_ _1"> </span>weakness<span class="_ _2"> </span>by<span class="_ _1"> </span>suppressing</div><div class="t m0 x8 h7 y25 ff2 fs3 fc0 sc0 ls0 ws0">responses<span class="_ _b"> </span>to<span class="_ _b"> </span>near-miss<span class="_ _b"> </span>or<span class="_ _b"> </span>distractor<span class="_ _b"> </span>patterns,<span class="_ _11"> </span>while<span class="_ _b"> </span>pre-</div><div class="t m0 x8 h7 y26 ff2 fs3 fc0 sc0 ls0 ws0">serving<span class="_ _1"> </span>strong<span class="_ _1"> </span>responses<span class="_ _1"> </span>to<span class="_ _1"> </span>the<span class="_ _1"> </span>target<span class="_ _7"> </span>pattern.<span class="_ _12"> </span>The<span class="_ _1"> </span>differ<span class="_ _6"></span>-</div><div class="t m0 x8 h7 y27 ff2 fs3 fc0 sc0 ls0 ws0">ences<span class="_ _2"> </span>among<span class="_ _2"> </span>these<span class="_ _2"> </span>&#64257;lters<span class="_ _c"> </span>lie<span class="_ _2"> </span>in<span class="_ _2"> </span>how<span class="_ _2"> </span>they<span class="_ _1"> </span>are<span class="_ _c"> </span>constructed</div><div class="t m0 x8 h7 y28 ff2 fs3 fc0 sc0 ls0 ws0">from<span class="_ _c"> </span>training<span class="_ _c"> </span>samples.<span class="_ _13"> </span>Examples<span class="_ _c"> </span>include<span class="_ _c"> </span>Synthetic<span class="_ _c"> </span>Dis-</div><div class="t m0 x8 h7 y29 ff2 fs3 fc0 sc0 ls0 ws0">criminant<span class="_ _10"> </span>Functions<span class="_ _11"> </span>(SDF)[5],<span class="_ _14"> </span>Minimum<span class="_ _10"> </span>V<span class="_ _a"></span>ariance<span class="_ _11"> </span>Syn-</div><div class="t m0 x8 h7 y2a ff2 fs3 fc0 sc0 ls0 ws0">thetic<span class="_ _1"> </span>Discriminant<span class="_ _1"> </span>Functions<span class="_ _1"> </span>(MVSDF)<span class="_ _1"> </span>&#64257;lters[15],<span class="_ _1"> </span>Mini-</div><div class="t m0 x9 h7 y2b ff2 fs3 fc0 sc0 ls0 ws0">mum<span class="_ _3"> </span>A<span class="_ _a"></span>verage<span class="_ _3"> </span>Correlation<span class="_ _3"> </span>Ener<span class="_ _6"></span>gy<span class="_ _3"> </span>(MACE)<span class="_ _8"> </span>&#64257;lters[8],<span class="_ _3"> </span>Opti-</div><div class="t m0 x9 h7 y2c ff2 fs3 fc0 sc0 ls0 ws0">mal<span class="_ _7"> </span>Tradeof<span class="_ _6"></span>f<span class="_ _7"> </span>Filters<span class="_ _1"> </span>(O<span class="_ _6"></span>TF)[12]<span class="_ _7"> </span>and<span class="_ _1"> </span>Unconstrained<span class="_ _7"> </span>MA<span class="_ _6"></span>CE</div><div class="t m0 x9 h7 y2d ff2 fs3 fc0 sc0 ls0 ws0">(UMA<span class="_ _6"></span>CE)[14].</div><div class="t m0 xa h7 y2e ff2 fs3 fc0 sc0 ls0 ws0">While<span class="_ _3"> </span>these<span class="_ _7"> </span>approaches<span class="_ _7"> </span>greatly<span class="_ _7"> </span>extend<span class="_ _3"> </span>the<span class="_ _7"> </span>performance</div><div class="t m0 x9 h7 y2f ff2 fs3 fc0 sc0 ls0 ws0">range<span class="_ _f"> </span>of<span class="_ _8"> </span>correlation<span class="_ _8"> </span>&#64257;lters,<span class="_ _8"> </span>there<span class="_ _8"> </span>is<span class="_ _8"> </span>still<span class="_ _8"> </span>further<span class="_ _8"> </span>room<span class="_ _8"> </span>for<span class="_ _f"> </span>im-</div><div class="t m0 x9 h7 y30 ff2 fs3 fc0 sc0 ls0 ws0">prov<span class="_ _6"></span>ement.<span class="_ _14"> </span>Speci&#64257;cally<span class="_ _a"></span>,<span class="_ _1"> </span>we<span class="_ _1"> </span>propose<span class="_ _7"> </span>a<span class="_ _1"> </span>ne<span class="_ _6"></span>w<span class="_ _1"> </span>class<span class="_ _7"> </span>of<span class="_ _1"> </span>&#64257;lters</div><div class="t m0 x9 h7 y31 ff2 fs3 fc0 sc0 ls0 ws0">called<span class="_ _f"> </span>A<span class="_ _6"></span>verage<span class="_ _f"> </span>of<span class="_ _8"> </span>Synthetic<span class="_ _8"> </span>Exact<span class="_ _f"> </span>Filters<span class="_ _8"> </span>(ASEF)<span class="_ _8"> </span>that<span class="_ _8"> </span>differ</div><div class="t m0 x9 h7 y32 ff2 fs3 fc0 sc0 ls0 ws0">from<span class="_ _7"> </span>these<span class="_ _7"> </span>prior<span class="_ _7"> </span>methods<span class="_ _7"> </span>in<span class="_ _7"> </span>two<span class="_ _7"> </span>important<span class="_ _1"> </span>respects.<span class="_ _b"> </span>First,</div><div class="t m0 x9 h7 y33 ff2 fs3 fc0 sc0 ls0 ws0">an<span class="_ _7"> </span>entire<span class="_ _1"> </span>correlation<span class="_ _7"> </span>response<span class="_ _7"> </span>surface<span class="_ _7"> </span>is<span class="_ _1"> </span>speci&#64257;ed<span class="_ _7"> </span>for<span class="_ _1"> </span>each</div><div class="t m0 x9 h7 y34 ff2 fs3 fc0 sc0 ls0 ws0">training<span class="_ _3"> </span>instance<span class="_ _3"> </span>during<span class="_ _3"> </span>&#64257;lter<span class="_ _3"> </span>construction.<span class="_ _c"> </span>Second,<span class="_ _7"> </span>the<span class="_ _3"> </span>re-</div><div class="t m0 x9 h7 y35 ff2 fs3 fc0 sc0 ls0 ws0">sulting<span class="_ _8"> </span>&#64257;lters,<span class="_ _3"> </span>one<span class="_ _8"> </span>per<span class="_ _3"> </span>training<span class="_ _8"> </span>image,<span class="_ _3"> </span>are<span class="_ _8"> </span>then<span class="_ _3"> </span>simply<span class="_ _8"> </span>aver<span class="_ _6"></span>-</div><div class="t m0 x9 h7 y36 ff2 fs3 fc0 sc0 ls0 ws0">aged.<span class="_ _3"> </span>The<span class="_ _8"> </span>resulting<span class="_ _8"> </span>&#64257;lters<span class="_ _8"> </span>are<span class="_ _8"> </span>less<span class="_ _f"> </span>susceptible<span class="_ _8"> </span>to<span class="_ _8"> </span>over<span class="_ _6"></span>-&#64257;tting</div><div class="t m0 x9 h7 y37 ff2 fs3 fc0 sc0 ls0 ws0">the<span class="_ _7"> </span>training<span class="_ _1"> </span>data<span class="_ _7"> </span>than<span class="_ _7"> </span>other<span class="_ _1"> </span>methods,<span class="_ _1"> </span>and<span class="_ _7"> </span>can<span class="_ _7"> </span>therefore<span class="_ _1"> </span>be</div><div class="t m0 x9 h7 y38 ff2 fs3 fc0 sc0 ls0 ws0">trained<span class="_ _1"> </span>ov<span class="_ _9"></span>er<span class="_ _1"> </span>larger<span class="_ _7"> </span>and<span class="_ _2"> </span>more<span class="_ _1"> </span>inclusiv<span class="_ _6"></span>e<span class="_ _1"> </span>training<span class="_ _1"> </span>sets.<span class="_ _12"> </span>As<span class="_ _1"> </span>a</div><div class="t m0 x9 h7 y39 ff2 fs3 fc0 sc0 ls0 ws0">result,<span class="_ _8"> </span>they<span class="_ _3"> </span>out<span class="_ _8"> </span>perform<span class="_ _8"> </span>previous<span class="_ _8"> </span>methods.</div><div class="t m0 xa h7 y3a ff2 fs3 fc0 sc0 ls0 ws0">On<span class="_ _2"> </span>the<span class="_ _2"> </span>task<span class="_ _2"> </span>of<span class="_ _2"> </span>&#64257;nding<span class="_ _2"> </span>eyes<span class="_ _2"> </span>in<span class="_ _2"> </span>face<span class="_ _2"> </span>images,<span class="_ _c"> </span>ASEF<span class="_ _2"> </span>out</div><div class="t m0 x9 h7 y3b ff2 fs3 fc0 sc0 ls0 ws0">performs<span class="_ _3"> </span>all<span class="_ _3"> </span>of<span class="_ _3"> </span>the<span class="_ _3"> </span>&#64257;lter<span class="_ _3"> </span>types<span class="_ _7"> </span>mentioned<span class="_ _3"> </span>above,<span class="_ _3"> </span>as<span class="_ _3"> </span>well<span class="_ _3"> </span>as</div><div class="t m0 x9 h7 y3c ff2 fs3 fc0 sc0 ls0 ws0">Haar-based<span class="_ _2"> </span>Cascade<span class="_ _2"> </span>Classi&#64257;ers<span class="_ _c"> </span>[17]<span class="_ _c"> </span>and<span class="_ _2"> </span>Gabor<span class="_ _c"> </span>Jet-based</div><div class="t m0 x9 h7 y3d ff2 fs3 fc0 sc0 ls0 ws0">methods[19].<span class="_ _13"> </span>This<span class="_ _e"> </span>new<span class="_ _2"> </span>family<span class="_ _c"> </span>of<span class="_ _e"> </span>correlation<span class="_ _c"> </span>&#64257;lters<span class="_ _e"> </span>thus</div><div class="t m0 x9 h7 y3e ff2 fs3 fc0 sc0 ls0 ws0">shows<span class="_ _3"> </span>state-of-the-art<span class="_ _3"> </span>performance<span class="_ _3"> </span>on<span class="_ _7"> </span>the<span class="_ _3"> </span>task<span class="_ _3"> </span>of<span class="_ _7"> </span>eye<span class="_ _3"> </span>&#64257;nd-</div><div class="t m0 x9 h7 y3f ff2 fs3 fc0 sc0 ls1 ws0">ing<span class="_ _1"> </span>and<span class="_ _1"> </span>we<span class="_ _1"> </span>anticipate<span class="_ _1"> </span>similar<span class="_ _1"> </span>performance<span class="_ _1"> </span>gains<span class="_ _1"> </span>on<span class="_ _1"> </span>many</div><div class="t m0 x9 h7 y40 ff2 fs3 fc0 sc0 ls0 ws0">challenging<span class="_ _8"> </span>object<span class="_ _3"> </span>detection<span class="_ _8"> </span>tasks.</div><div class="t m0 x9 h5 y41 ff1 fs1 fc0 sc0 ls0 ws0">2.<span class="_ _3"> </span>Background</div><div class="t m0 xa h7 y42 ff2 fs3 fc0 sc0 ls0 ws0">Accurate<span class="_ _8"> </span>registration<span class="_ _3"> </span>of<span class="_ _8"> </span>face<span class="_ _8"> </span>images<span class="_ _3"> </span>is<span class="_ _8"> </span>an<span class="_ _8"> </span>important<span class="_ _3"> </span>&#64257;rst</div><div class="t m0 x9 h7 y43 ff2 fs3 fc0 sc0 ls0 ws0">step<span class="_ _8"> </span>in<span class="_ _8"> </span>face<span class="_ _8"> </span>recognition,<span class="_ _3"> </span>and<span class="_ _8"> </span>one<span class="_ _8"> </span>common<span class="_ _8"> </span>way<span class="_ _3"> </span>of<span class="_ _f"> </span>establish-</div><div class="t m0 x9 h7 y44 ff2 fs3 fc0 sc0 ls0 ws0">ing<span class="_ _f"> </span>face<span class="_ _f"> </span>registration<span class="_ _f"> </span>is<span class="_ _f"> </span>by<span class="_ _8"> </span>&#64257;nding<span class="_ _f"> </span>eyes.<span class="_ _3"> </span>As<span class="_ _8"> </span>noted<span class="_ _f"> </span>below<span class="_ _6"></span>,<span class="_ _f"> </span>cor-</div><div class="t m0 x9 h7 y45 ff2 fs3 fc0 sc0 ls0 ws0">relation<span class="_ _7"> </span>&#64257;lters<span class="_ _7"> </span>hav<span class="_ _6"></span>e<span class="_ _7"> </span>been<span class="_ _7"> </span>applied<span class="_ _7"> </span>to<span class="_ _7"> </span>eye<span class="_ _3"> </span>&#64257;nding<span class="_ _7"> </span>before<span class="_ _7"> </span>[3].</div><div class="t m0 x9 h7 y46 ff2 fs3 fc0 sc0 ls0 ws0">This<span class="_ _3"> </span>section<span class="_ _7"> </span>re<span class="_ _9"></span>views<span class="_ _3"> </span>the<span class="_ _3"> </span>eye<span class="_ _3"> </span>&#64257;nding<span class="_ _3"> </span>task<span class="_ _7"> </span>and<span class="_ _3"> </span>past<span class="_ _3"> </span>work<span class="_ _7"> </span>on</div><div class="t m0 x9 h7 y47 ff2 fs3 fc0 sc0 ls0 ws0">synthetic<span class="_ _2"> </span>correlation<span class="_ _2"> </span>&#64257;lters.<span class="_ _15"> </span>Emphasis<span class="_ _2"> </span>is<span class="_ _2"> </span>placed<span class="_ _c"> </span>on<span class="_ _2"> </span>algo-</div><div class="t m0 x9 h7 y48 ff2 fs3 fc0 sc0 ls0 ws0">rithms<span class="_ _3"> </span>that<span class="_ _7"> </span>appear<span class="_ _7"> </span>in<span class="_ _3"> </span>the<span class="_ _7"> </span>direct<span class="_ _3"> </span>comparisons<span class="_ _7"> </span>with<span class="_ _7"> </span>ASEF<span class="_ _3"> </span>in</div><div class="t m0 x9 h7 y49 ff2 fs3 fc0 sc0 ls0 ws0">Section<span class="_ _8"> </span>4.</div><div class="t m0 x9 h8 y4a ff1 fs4 fc0 sc0 ls0 ws0">2.1.<span class="_ _3"> </span>Eye<span class="_ _3"> </span>Finding/Localization</div><div class="t m0 xa h7 y4b ff2 fs3 fc0 sc0 ls0 ws0">Generally<span class="_ _6"></span>,<span class="_ _7"> </span>eye<span class="_ _7"> </span>&#64257;nding<span class="_ _7"> </span>algorithms<span class="_ _7"> </span>return<span class="_ _1"> </span>the<span class="_ _7"> </span>pixel<span class="_ _3"> </span>coor-</div><div class="t m0 x9 h7 y4c ff2 fs3 fc0 sc0 ls0 ws0">dinates<span class="_ _3"> </span>of<span class="_ _3"> </span>the<span class="_ _3"> </span>center<span class="_ _3"> </span>of<span class="_ _3"> </span>the<span class="_ _3"> </span>left<span class="_ _3"> </span>and<span class="_ _7"> </span>right<span class="_ _3"> </span>eyes<span class="_ _3"> </span>in<span class="_ _3"> </span>frontal,<span class="_ _3"> </span>or</div><div class="t m0 x9 h7 y4d ff2 fs3 fc0 sc0 ls0 ws0">near<span class="_ _7"> </span>frontal,<span class="_ _1"> </span>face<span class="_ _7"> </span>images.<span class="_ _11"> </span>T<span class="_ _6"></span>o<span class="_ _7"> </span>be<span class="_ _7"> </span>counted<span class="_ _1"> </span>as<span class="_ _7"> </span>a<span class="_ _7"> </span>success,<span class="_ _1"> </span>the</div><div class="t m0 x9 h7 y4e ff2 fs3 fc0 sc0 ls0 ws0">algorithm<span class="_ _3"> </span>must<span class="_ _7"> </span>return<span class="_ _7"> </span>the<span class="_ _3"> </span>true<span class="_ _7"> </span>eye<span class="_ _3"> </span>location<span class="_ _7"> </span>to<span class="_ _7"> </span>within<span class="_ _3"> </span>some</div><div class="t m0 x9 h7 y4f ff2 fs3 fc0 sc0 ls0 ws0">tolerance,<span class="_ _f"> </span>typically<span class="_ _8"> </span>measured<span class="_ _f"> </span>as<span class="_ _f"> </span>a<span class="_ _8"> </span>fraction<span class="_ _f"> </span>of<span class="_ _8"> </span>the<span class="_ _f"> </span>interocular</div><div class="t m0 xb h7 y50 ff2 fs3 fc0 sc0 ls0 ws0"><span class="fc1 sc0">1</span></div><div class="t m0 xc h9 y51 ff5 fs5 fc0 sc0 ls0 ws0">2105<span class="_ _16"></span>978-1-4244-3991-1/09/$25.00 &#169;2009 IEEE</div><div class="t m0 xd ha y52 ff6 fs6 fc0 sc0 ls0 ws0">Authorized licensed use limited to: SOUTHWEST JIAOTONG UNIVERSITY. Downloaded on December 5, 2009 at 05:10 from IEEE Xplore. Restrictions apply. </div></div><div class="pi" data-data='{"ctm":[1.568627,0.000000,0.000000,1.568627,0.000000,0.000000]}'></div></div> </body> </html>
评论
    相关推荐
    • 人脸识别.rar
      人脸识别程序, 人脸识别,特定人脸识别,人脸识别,特定人脸识别。。。。。。
    • 人脸识别.rar
      使用百度智能云的api端口,实现人脸检测
    • 人脸识别
      适用于人脸识别终端、人脸支付终端及安防类终端产品,产品体积小巧,便于嵌入整机设备,如:智慧闸机、人脸检验一体机、新零售智能终端、O2O 智能设备、机器人设备、AI 设备等
    • 虹软人脸识别delphi
      1.上虹软官网开通申请免费SDK 3.0,申请到appid,需要在线激活。 2.摄像头用的组件是devexpress19.1.2
    • 人脸识别资源
      人脸识别资源人脸识别资源人脸识别资源人脸识别资源人脸识别资源人脸识别资源
    • 基于opencv的人脸识别
      毕业设计,人脸识别门禁,学习,可以运行,python
    • 人脸识别数据库
      广义的人脸识别实际包括构建人脸识别系统的一系列相关技术,包括人脸图像采集、人脸定位、人脸识别预处理、身份确认以及身份查找等;而狭义的人脸识别特指通过人脸进行身份确认或者身份查找的技术或系统。[1] 人脸...
    • 人脸识别
      人脸识别与虹膜识别、指纹识别、步态识别等其它生物特征识别技术相比,具有自然、便捷、用户体验友好等独特优势,因而受到了学术界和工业界的广泛关注.
    • 模式识别人脸识别
      本实验流程说明书描述了怎样利用一个基于神经网络的自动人脸识别模型来进行人脸图片识别的流程,以及试验环境和参数输入规范
    • VB人脸识别技术
      VB人脸识别技术 模式识别技术应用于脸部图像特征对比 图像输入 直方图,阀值,边缘提取,二值化