1045 |
% subjective preference for text displays and empirical measures of their |
% subjective preference for text displays and empirical measures of their |
1046 |
% readability." |
% readability." |
1047 |
|
|
1048 |
|
It can be argued that |
1049 |
|
the backgrounds clutter the display |
1050 |
|
visually, making the user interface more confusing, |
1051 |
|
and reduce text readability. |
1052 |
|
Indeed, one of the most difficult aspects of the work was making the random |
1053 |
|
color selection produce acceptable results. |
1054 |
|
However, by tuning the color selection and the gamma |
1055 |
|
correction of the display, we were able to (in our opinion) avoid |
1056 |
|
the above problems. |
1057 |
|
It is important that the colors chosen are light and that the palettes |
1058 |
|
have a relatively small range of colors. |
1059 |
|
Also, text readability on the generated backgrounds depends |
1060 |
|
strongly on the text scale; making it easy for the user to zoom |
1061 |
|
fluidly in and out helps. |
1062 |
|
|
1063 |
|
We have not tried to attain infinite zoomability\cite{furnas00infinity} |
1064 |
|
with the current implementation, |
1065 |
|
but only the more modest goal of zooming within a range |
1066 |
|
that would be reasonable for a single PDF document, |
1067 |
|
i.e., approximately 100-fold |
1068 |
|
difference between mininum and maximum zoom. |
1069 |
|
% This is in line with our |
1070 |
|
% intended use, since it would be unreasonable |
1071 |
|
% to expect a texture to be recognizable |
1072 |
|
% if a sub-pixel area were zoomed to the full screen. |
1073 |
|
It could be possible\cite{furnas00infinity} |
1074 |
|
to make the unique background look similar |
1075 |
|
at different scales, but |
1076 |
|
this would remove the use of the texture as a cue of scale. |
1077 |
|
Our |
1078 |
|
nonlinear use of the register combiners |
1079 |
|
does have some ill effects when zooming the texture out |
1080 |
|
to a very small scale: mipmapping will not give the correct |
1081 |
|
average color value. |
1082 |
|
It may be possible to alleviate this by modeling the texture mathematically |
1083 |
|
and calculating the correct average and placing corrective terms to the |
1084 |
|
equations. |
1085 |
|
However, in the intended zooming range |
1086 |
|
the current system is quite satisfactory. |
1087 |
|
|
1088 |
|
% It is important that the background can be zoomed |
1089 |
|
% to different resolutions. |
1090 |
|
|
1091 |
|
% leads to aliasing: .... modeling textures mathematically , ... |
1092 |
|
|
1093 |
TJL |
TJL |
1094 |
|
|
1095 |
The most commonly asked question about this work concerns |
The most commonly asked question about this work concerns |
1115 |
|
|
1116 |
JVK |
JVK |
1117 |
|
|
|
There has been |
|
|
lot of texture perception work on texture discrimination. |
|
|
However, in our application texture discrimination is not as |
|
|
much of an issue as memorizability and recognizability of |
|
|
previously seen textures. |
|
|
Furthermore, our textures are on a higher level, |
|
|
more like complete pictures than the usually studied microstructure. |
|
|
|
|
|
|
|
|
XXX |
|
|
|
|
|
%\subsection{Texture perception} |
|
|
|
|
1118 |
Psychophysical studies on texture perception have mostly concentrated |
Psychophysical studies on texture perception have mostly concentrated |
1119 |
on \emph{texture discrimination}, the ability of human observers to |
on \emph{texture discrimination}\cite{julesz62visualpattern}, |
1120 |
discriminate pairs of textures. |
the ability of human observers to discriminate pairs of textures. |
1121 |
The term is often used interchangably with \emph{texture segregation}, |
%The term is often used interchangably with \emph{texture segregation}, |
1122 |
the more specific task of finding the border between differently textured |
%the more specific task of finding the border between differently textured |
1123 |
areas (different phases of local characteristics at the |
%areas (different phases of local characteristics at the |
1124 |
border can segregate otherwise indiscriminable textures). |
%border can segregate otherwise indiscriminable textures). |
1125 |
|
% |
1126 |
First experiments on computer-generated, unnatural textures in the 60s |
%First experiments on computer-generated, unnatural textures in the 60s |
1127 |
\cite{julesz62visualpattern} led to proposals of discrimination models |
%\cite{julesz62visualpattern} led to proposals of discrimination models |
1128 |
based on the $N$th-order statistics of textures |
%based on the $N$th-order statistics of textures |
1129 |
|
%(the joint distributions of the values at the corners of a randomly |
1130 |
|
%placed (translated) $N$-gon for all different $N$-gons). |
1131 |
|
%%and connectivity structures of certain micropatterns. |
1132 |
|
% |
1133 |
|
First discrimination models were based |
1134 |
|
on the $N$th-order statistics of textures |
1135 |
(the joint distributions of the values at the corners of a randomly |
(the joint distributions of the values at the corners of a randomly |
1136 |
placed (translated) $N$-gon for all different $N$-gons). |
placed (translated) $N$-gon for all different $N$-gons). |
1137 |
%and connectivity structures of certain micropatterns. |
However, the order of similarity in the statistics did not |
1138 |
|
consistently explain discrimination performance, and certain |
1139 |
|
distinctive local features were conjectured. |
1140 |
|
|
1141 |
Attempt to explain texture discrimination by the densities of textons |
Julesz\cite{julesz81textons} proposed that discrimination could be explained |
1142 |
\cite{julesz81textons}, fundamental texture elements, such as |
by the densities of textons, fundamental texture elements, such as |
1143 |
elongated blobs, line terminators, line crossings, etc. |
elongated blobs, line terminators, line crossings, etc. |
1144 |
However, the textons are hard to define formally. |
However, the textons are hard to define formally. |
1145 |
|
|
1146 |
Much simpler filtering-based models can explain texture discrimination |
Much simpler filtering-based models can explain texture discrimination |
1147 |
just as well \cite{bergen88earlyvision}. |
just as well \cite{bergen88earlyvision}. |
1148 |
Essentially a bank of linear filters is applied to the texture followed |
In this approach, a bank of linear filters is applied to the texture followed |
1149 |
by a nonlinearity and then another set of filters. |
by a nonlinearity and then another set of filters to extract densities |
1150 |
In \cite{heeger95pyramid}, new textures with appearance similar |
of features (see, e.g., \cite{heeger95pyramid} for an application). |
1151 |
to a given texture are created by matching certain histograms |
%In \cite{heeger95pyramid}, new textures with appearance similar |
1152 |
of filter responses. |
%to a given texture are created by matching certain histograms |
1153 |
|
%of filter responses. |
1154 |
|
|
1155 |
Mapping texture appearance to an Euclidian texture space |
In our application texture discrimination is not as |
1156 |
(see \cite{gurnsey01texturespace} and the references therein): |
much of an issue as memorizability and recognizability of |
1157 |
in the reported experiments, three dimensions have been sufficient |
previously seen textures. |
1158 |
to explain most of the variation in the similarity judgements for |
Furthermore, in most texture perception work |
1159 |
artificial textures. |
texture is considered as the output of a stochastic |
|
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 |
|
1160 |
process that produces certain repeating features. |
process that produces certain repeating features. |
|
Different samples from the process are considered as the same texture. |
|
1161 |
The textures created by our algorithm, although repeating, are more like |
The textures created by our algorithm, although repeating, are more like |
1162 |
complete images rather than microstructure. |
complete images than statistical microstructure. |
1163 |
Therefore, higher level processes of vision are also involved |
Therefore, higher level processes of vision are also involved |
1164 |
in the perception and recognition. |
in the perception and recognition. |
1165 |
|
|
|
theories of structural object perception |
|
|
(see, e.g., Biederman\cite{biederman87}) |
|
|
|
|
|
|
|
|
XXX |
|
|
|
|
1166 |
Experiments on black-and-white %(faces,) |
Experiments on black-and-white %(faces,) |
1167 |
ink blots, and snow crystals |
ink blots, and snow crystals |
1168 |
\cite{goldstein71visualrecognition} show that |
\cite{goldstein71visualrecognition} show that |
1201 |
|
|
1202 |
XXX: refs? |
XXX: refs? |
1203 |
|
|
1204 |
XXX: texture set dimensionality studies? |
Mapping texture appearance to an Euclidian texture space |
1205 |
|
(see \cite{gurnsey01texturespace} and the references therein): |
1206 |
|
in the reported experiments, three dimensions have been sufficient |
1207 |
|
to explain most of the variation in the similarity judgements for |
1208 |
|
artificial textures. |
1209 |
|
However, the texture stimuli have been somewhat simple |
1210 |
|
(no color, lack of frequency-band interaction, etc.). |
1211 |
|
For some natural texture sets (see, e.g., \cite{rao96texturenaming}), |
1212 |
|
three dimensions have also been |
1213 |
|
sufficient, but often semantic connections cause the |
1214 |
|
similarity to be context-dependant, making it hard to assess the |
1215 |
|
dimensionality. |
1216 |
|
% XXX: this is something we should experiment with our textures |
1217 |
|
|
1218 |
|
XXX: reviews |
1219 |
|
|
1220 |
|
XXX: physiological knowledge of visual perception |
1221 |
|
(see, e.g.,~Bruce et al\cite{bruce96visualperception}); |
1222 |
|
|
1223 |
|
theories of structural object perception |
1224 |
|
(see, e.g., Biederman\cite{biederman87}) |
1225 |
|
|
1226 |
|
|
1227 |
|
XXX |
1228 |
|
|
1229 |
%\section{Software availability} |
%\section{Software availability} |
1230 |
|
|
1270 |
to give, e.g., all academic articles stored on a user's hard |
to give, e.g., all academic articles stored on a user's hard |
1271 |
drive their own background. |
drive their own background. |
1272 |
|
|
1273 |
\subsection{Problems} |
%\subsection{Further work} |
|
|
|
|
It can be argued that |
|
|
the backgrounds clutter the display |
|
|
visually, making the user interface more confusing, |
|
|
and reduce text readability. |
|
|
Indeed, one of the most difficult aspects of the work was making the random |
|
|
color selection produce acceptable results. |
|
|
However, by tuning the color selection and the gamma |
|
|
correction of the display, we were able to (in our opinion) avoid |
|
|
the above problems. |
|
|
It is important that the colors chosen are light and that the palettes |
|
|
have a relatively small range of colors. |
|
|
Also, text readability on the generated backgrounds depends |
|
|
strongly on the text scale; making it easy for the user to zoom |
|
|
fluidly in and out helps. |
|
|
|
|
|
We have not tried to attain infinite zoomability\cite{furnas00infinity} |
|
|
with the current implementation, |
|
|
but only the more modest goal of zooming within a range |
|
|
that would be reasonable for a single PDF document, |
|
|
i.e., approximately 100-fold |
|
|
difference between mininum and maximum zoom. |
|
|
% This is in line with our |
|
|
% intended use, since it would be unreasonable |
|
|
% to expect a texture to be recognizable |
|
|
% if a sub-pixel area were zoomed to the full screen. |
|
|
It could be possible\cite{furnas00infinity} |
|
|
to make the unique background look similar |
|
|
at different scales, but |
|
|
this would remove the use of the texture as a cue of scale. |
|
|
Our |
|
|
nonlinear use of the register combiners |
|
|
does have some ill effects when zooming the texture out |
|
|
to a very small scale: mipmapping will not give the correct |
|
|
average color value. |
|
|
It may be possible to alleviate this by modeling the texture mathematically |
|
|
and calculating the correct average and placing corrective terms to the |
|
|
equations. |
|
|
However, in the intended zooming range |
|
|
the current system is quite satisfactory. |
|
|
|
|
|
% It is important that the background can be zoomed |
|
|
% to different resolutions. |
|
|
|
|
|
% leads to aliasing: .... modeling textures mathematically , ... |
|
|
|
|
|
\subsection{Further work} |
|
1274 |
|
|
1275 |
So far, we have concentrated mostly on low-end hardware, and |
So far, we have concentrated mostly on low-end hardware, and |
1276 |
have not even tapped the full potential of the NV25 architecture. |
have not even tapped the full potential of the NV25 architecture. |