texture_analysis_using_gabor_filter

texture_analysis_using_gabor_filter

ID:39448404

大小:1.99 MB

页数:38页

时间:2019-07-03

texture_analysis_using_gabor_filter_第1页
texture_analysis_using_gabor_filter_第2页
texture_analysis_using_gabor_filter_第3页
texture_analysis_using_gabor_filter_第4页
texture_analysis_using_gabor_filter_第5页
资源描述:

《texture_analysis_using_gabor_filter》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、TEXTUREANALYSISUSINGGABORFILTERSTextureTypes•DefinitionofTexture•TexturetypesSyntheticNaturalStochasticTextureDefinitionTexture:theregularrepetitionofanelementorpatternonasurface.•Purposeoftextureanalysis:–Toidentifydifferenttexturedandnon-texturedregionsinanimage.–Tocl

2、assify/segmentdifferenttextureregionsinanimage.–Toextractboundariesbetweenmajortextureregions.–Todescribethetexelunit.–3-DshapefromtextureRef:[Forsyth2003,Raghu95]TextureRegions&EdgesIITMadrasR1R2R1R2R3R4R3R4TexturedimageImagehistogramsVPLABInputImageFILTERINGSMOOTHINGC

3、LASSIFIERSegmentedImageFlow-chartofatypicalmethodoftextureclassificationSyntheticTextureImageHorizontalintensityFilteredprofileoutputNonlinearSmoothingtransformSegmentedimageProcessingofTexture-likeImages2-DGaborFilter1⎡−1x2y2⎤f(x,y,ω,θ,σx,σy)=exp(()+())+jω(xcosθ+ysinθ)⎢⎥2πσxσy⎣2

4、σxσy⎦AtypicalGaussianAtypicalGaborfilterwithfilterwithσ=30σ=30,ω=3.14andθ=45°Gaborfilterswithdifferentcombinationsofspatialwidthσ,frequencyωandorientationθ.2-DGaborfilter:1⎡−1x2y2⎤f(x,y,ω,θ,σx,σy)=exp(()+())+jω(xcosθ+ysinθ)2πσxσy⎢⎣2σxσy⎥⎦whereσisthespatialspreadωisthefr

5、equencyθistheorientation1-DGaborfilter:1−x2f(x,ω,σ)=exp(+jωx)22πσ2σ21-DGaussianfunction:1⎛−x⎞g(x)=exp⎜⎟⎜2⎟2πσ1⎝2σ⎠ProcessingofTexture-likeImages1-DGaborFilter1-DGaussianFilter21−x2f(x,ω,σ)=exp(+jωx)1⎛−x⎞22g(x)=exp⎜⎟2πσσ⎜2⎟2πσ1⎝2σ⎠EvenComponentOddComponent2222g(x,y)=Kexp

6、(−π(a(x−x)+b(y−y)))0θ0θexp(j(2πF(xcosω+ysinω)+P)000•K:ScalesthemagnitudeoftheGaussianenvelop.•(a,b):ScalethetwoaxisoftheGaussianenvelop.•θ:(Rotation)angleoftheGaussianenvelop.•(x,y):Locationofthepeakofthe00Gaussianenvelop.•(u,v):Spatialfrequenciesofthe00sinusoidalcarrierinCartesi

7、ancoordinates.Itcanalsobeexpressedinpolarcoordinatesas(F,ω).00g(x,y)=•P:Phaseofthesinusoidalcarrier.2222Kexp(−π(a(x−x)+b(y−y))+j(2π(ux+vy)+P)0θ0θ00Asymmetric2-DGaussianfunctionAsymmetricalGaussianof128×128pixels.Theparametersareasfollows:x=y=0;a=1/50pixels;b=1/40pixels;θ=−45deg.0

8、0TherealandimaginarypartsofacomplexGabor

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。
相关文章
更多
相关标签