/[gzz]/manuscripts/Paper/paper.tex
ViewVC logotype

Diff of /manuscripts/Paper/paper.tex

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.30 by jvk, Tue Mar 18 09:23:34 2003 UTC revision 1.31 by jvk, Tue Mar 18 14:06:47 2003 UTC
# Line 26  Janne V. Kujala and Tuomas J. Lukka\\ Line 26  Janne V. Kujala and Tuomas J. Lukka\\
26    University of Jyv\"askyl\"a, PO.~Box~35\\    University of Jyv\"askyl\"a, PO.~Box~35\\
27    FIN-40351~Jyv\"askyl\"a\\    FIN-40351~Jyv\"askyl\"a\\
28    Finland\\    Finland\\
29    lukka@iki.fi,\ \  jvk@iki.fi    jvk@iki.fi,\ \  lukka@iki.fi
30  }  }
31    
32  \begin{document}  \begin{document}
# Line 244  other textures as a starting point Line 244  other textures as a starting point
244  and perceptually, for visualizing surface orientation\cite{schweitzer83texturing}, scalar or vector fields\cite{ware95texture}, and  and perceptually, for visualizing surface orientation\cite{schweitzer83texturing}, scalar or vector fields\cite{ware95texture}, and
245  surface shape\cite{interrante97illustrating}.  surface shape\cite{interrante97illustrating}.
246    
 % In this article, we apply texture shading to synthesize a large number  
 % of unique textures for distinguishing virtual objects.  
   
 \subsection{Texture perception}  
   
 Psychological studies on texture perception have mostly concentrated  
 on \emph{texture discrimination}, the ability of human observers to  
 discriminate pairs of textures.    
 The term is often used interchangably with \emph{texture segregation},  
 the more specific task of finding the border between differently textured  
 areas (different phases of local characteristics at the  
 border can segregate otherwise indiscriminable textures).  
   
 First experiments on computer-generated, unnatural textures in the 60s  
 \cite{julesz62visualpattern} led to proposals of discrimination models  
 based on the $N$th-order statistics of textures  
 (the joint distributions of the values at the corners of a randomly  
 placed (translated) $N$-gon for all different $N$-gons).  
 %and connectivity structures of certain micropatterns.  
   
247  Statistical modeling of textures as samples from a probability  Statistical modeling of textures as samples from a probability
248  distribution on a random field as already seen in \cite{julesz62visualpattern}  distribution on a random field as already seen in \cite{julesz62visualpattern}
249  in a simple form.  in a simple form.
# Line 272  The most popualar computational approach Line 252  The most popualar computational approach
252  depends only on the values of its neighborhood (local characteristics).  depends only on the values of its neighborhood (local characteristics).
253  XXX: resolution-dependency?  XXX: resolution-dependency?
254    
255  Attempt to explain texture discrimination by the densities of textons  % In this article, we apply texture shading to synthesize a large number
256  \cite{julesz81textons}, fundamental texture elements, such as  % of unique textures for distinguishing virtual objects.
 elongated blobs, line terminators, line crossings, etc.    
 However, the textons are hard to define formally.  
   
 Much simpler filtering-based models can explain texture discrimination  
 just as well \cite{bergen88earlyvision}.  
 Essentially a bank of linear filters is applied to the texture followed  
 by a nonlinearity and then another set of filters.  
 In \cite{heeger95pyramid}, new textures with appearance similar  
 to a given texture are created by matching certain histograms  
 of filter responses.  
   
 Mapping texture appearance to an Euclidian texture space  
 (see \cite{gurnsey01texturespace} and the references therein):  
 in the reported experiments, three dimensions have been sufficient  
 to explain most of the variation in the similarity judgements for  
 artificial textures.  
 However, the texture stimuli have been somewhat simple  
 (no color, lack of frequency-band interaction, etc.).  
 For some natural texture sets (see, e.g., \cite{rao96texturenaming}),  
 three dimensions have also been  
 sufficient, but often semantic connections cause the  
 similarity to be context-dependant, making it hard to assess the  
 dimensionality.  
 % XXX: this is something we should experiment with our textures  
   
 XXX: reviews  
   
 XXX: physiological knowledge of visual perception  
 (see, e.g.,~Bruce et al\cite{bruce96visualperception});  
   
 XXX: in most work, texture is considered as the output of a stochastic  
 process that produces certain repeating features.  
 Different samples from the process are considered as the same texture.  
 The textures created by our algorithm, although repeating, are more like  
 complete images rather than microstructure.  
 Therefore, higher level processes of vision are also involved  
 in the perception and recognition.  
   
 theories of structural object perception  
 (see, e.g., Biederman\cite{biederman87})  
257    
258  \subsection{Focus+Context views}  \subsection{Focus+Context views}
259    
# Line 470  Now, Line 410  Now,
410    
411  \section{Unique Background Textures}  \section{Unique Background Textures}
412    
413    %XXX: shorten by one half column
414    
415  XXX: - visual discrimination experiments: \cite{julesz62visualpattern}  XXX: - visual discrimination experiments: \cite{julesz62visualpattern}
416    
417  XXX: simple models (filtering) can have good explanatory power  XXX: simple models (filtering) can have good explanatory power
# Line 1106  On a darker background, this approach wo Line 1048  On a darker background, this approach wo
1048    
1049  \subsection{Recognizability and memorizability}  \subsection{Recognizability and memorizability}
1050    
1051    % XXX: shorten by half
1052    
1053  JVK  JVK
1054    
# Line 1117  previously seen textures. Line 1060  previously seen textures.
1060  Furthermore, our textures are on a higher level,  Furthermore, our textures are on a higher level,
1061  more like complete pictures than the usually studied microstructure.  more like complete pictures than the usually studied microstructure.
1062    
1063    
1064    XXX
1065    
1066    %\subsection{Texture perception}
1067    
1068    Psychophysical studies on texture perception have mostly concentrated
1069    on \emph{texture discrimination}, the ability of human observers to
1070    discriminate pairs of textures.  
1071    The term is often used interchangably with \emph{texture segregation},
1072    the more specific task of finding the border between differently textured
1073    areas (different phases of local characteristics at the
1074    border can segregate otherwise indiscriminable textures).
1075    
1076    First experiments on computer-generated, unnatural textures in the 60s
1077    \cite{julesz62visualpattern} led to proposals of discrimination models
1078    based on the $N$th-order statistics of textures
1079    (the joint distributions of the values at the corners of a randomly
1080    placed (translated) $N$-gon for all different $N$-gons).
1081    %and connectivity structures of certain micropatterns.
1082    
1083    Attempt to explain texture discrimination by the densities of textons
1084    \cite{julesz81textons}, fundamental texture elements, such as
1085    elongated blobs, line terminators, line crossings, etc.  
1086    However, the textons are hard to define formally.
1087    
1088    Much simpler filtering-based models can explain texture discrimination
1089    just as well \cite{bergen88earlyvision}.
1090    Essentially a bank of linear filters is applied to the texture followed
1091    by a nonlinearity and then another set of filters.
1092    In \cite{heeger95pyramid}, new textures with appearance similar
1093    to a given texture are created by matching certain histograms
1094    of filter responses.
1095    
1096    Mapping texture appearance to an Euclidian texture space
1097    (see \cite{gurnsey01texturespace} and the references therein):
1098    in the reported experiments, three dimensions have been sufficient
1099    to explain most of the variation in the similarity judgements for
1100    artificial textures.
1101    However, the texture stimuli have been somewhat simple
1102    (no color, lack of frequency-band interaction, etc.).
1103    For some natural texture sets (see, e.g., \cite{rao96texturenaming}),
1104    three dimensions have also been
1105    sufficient, but often semantic connections cause the
1106    similarity to be context-dependant, making it hard to assess the
1107    dimensionality.
1108    % XXX: this is something we should experiment with our textures
1109    
1110    XXX: reviews
1111    
1112    XXX: physiological knowledge of visual perception
1113    (see, e.g.,~Bruce et al\cite{bruce96visualperception});
1114    
1115    XXX: in most work, texture is considered as the output of a stochastic
1116    process that produces certain repeating features.
1117    Different samples from the process are considered as the same texture.
1118    The textures created by our algorithm, although repeating, are more like
1119    complete images rather than microstructure.
1120    Therefore, higher level processes of vision are also involved
1121    in the perception and recognition.
1122    
1123    theories of structural object perception
1124    (see, e.g., Biederman\cite{biederman87})
1125    
1126    
1127    XXX
1128    
1129  Experiments on black-and-white %(faces,)  Experiments on black-and-white %(faces,)
1130  ink blots, and snow crystals  ink blots, and snow crystals
1131  \cite{goldstein71visualrecognition} show that  \cite{goldstein71visualrecognition} show that

Legend:
Removed from v.1.30  
changed lines
  Added in v.1.31

savannah-hackers-public@gnu.org
ViewVC Help
Powered by ViewVC 1.1.26