73 |
|
|
74 |
|
|
75 |
%FIXME: more concise and exact explanations in some sections |
%FIXME: more concise and exact explanations in some sections |
76 |
%FIXME: what are this paper's spefific contributions? (e.g., in Sec. 4) |
%FIXME: what are this paper's specific contributions? (e.g., in Sec. 4) |
77 |
|
|
78 |
%NO NEED FOR PERFECT RECALL - JUST A CUE |
%NO NEED FOR PERFECT RECALL - JUST A CUE |
79 |
|
|
83 |
We present a perceptually designed hardware-accelerated |
We present a perceptually designed hardware-accelerated |
84 |
algorithm for generating unique background textures for distinguishing |
algorithm for generating unique background textures for distinguishing |
85 |
documents. |
documents. |
86 |
To be recongizable, |
To be recognizable, |
87 |
the texture should produce a random feature vector in the brain |
the texture should produce a random feature vector in the brain |
88 |
after visual feature extraction. |
after visual feature extraction. |
89 |
|
|
280 |
|
|
281 |
%Textures have also been modeled statistically, |
%Textures have also been modeled statistically, |
282 |
%as samples from a probability distribution on a random field. |
%as samples from a probability distribution on a random field. |
283 |
%The most popualar computational approach is Markov random fields |
%The most popular computational approach is Markov random fields |
284 |
%\cite{cross83markov,geman84stochastic}, |
%\cite{cross83markov,geman84stochastic}, |
285 |
%where the value of each pixel |
%where the value of each pixel |
286 |
%depends only on the values of its neighborhood (local characteristics). |
%depends only on the values of its neighborhood (local characteristics). |
336 |
For some natural texture sets, |
For some natural texture sets, |
337 |
three dimensions have also been |
three dimensions have also been |
338 |
sufficient\cite{rao96texturenaming}, but often semantic connections cause the |
sufficient\cite{rao96texturenaming}, but often semantic connections cause the |
339 |
similarity to be context-dependant, making it hard to assess the |
similarity to be context-dependent, making it hard to assess the |
340 |
dimensionality. |
dimensionality. |
341 |
%% XXX: this is something we should experiment with our textures |
%% XXX: this is something we should experiment with our textures |
342 |
|
|
350 |
%\emph{visual texture discrimination}\cite{julesz62visualpattern}, |
%\emph{visual texture discrimination}\cite{julesz62visualpattern}, |
351 |
%the ability of human observers to effortlessly discriminate |
%the ability of human observers to effortlessly discriminate |
352 |
%pairs of certain textures (see Bergen\cite{bergen91theories} for a review). |
%pairs of certain textures (see Bergen\cite{bergen91theories} for a review). |
353 |
%%The term is often used interchangably with \emph{texture segregation}, |
%%The term is often used interchangeably with \emph{texture segregation}, |
354 |
%%the more specific task of finding the border between differently textured |
%%the more specific task of finding the border between differently textured |
355 |
%%areas (different phases of local characteristics at the |
%%areas (different phases of local characteristics at the |
356 |
%%border can segregate otherwise indiscriminable textures). |
%%border can segregate otherwise indiscriminable textures). |
399 |
%For some natural texture sets, |
%For some natural texture sets, |
400 |
%three dimensions have also been |
%three dimensions have also been |
401 |
%sufficient \cite{rao96texturenaming}, but often semantic connections cause the |
%sufficient \cite{rao96texturenaming}, but often semantic connections cause the |
402 |
%similarity to be context-dependant, making it hard to assess the |
%similarity to be context-dependent, making it hard to assess the |
403 |
%dimensionality. |
%dimensionality. |
404 |
%% XXX: this is something we should experiment with our textures |
%% XXX: this is something we should experiment with our textures |
405 |
|
|
516 |
hypertext. BuoyOING is a logical step from the earlier work |
hypertext. BuoyOING is a logical step from the earlier work |
517 |
on Fluid Links\cite{zellweger98fluid}, hypercept animations\cite{milgram99hypercept}, |
on Fluid Links\cite{zellweger98fluid}, hypercept animations\cite{milgram99hypercept}, |
518 |
and transpointing windows\cite{ted-xanalogical-structure-needed}. |
and transpointing windows\cite{ted-xanalogical-structure-needed}. |
519 |
Buoyoing has been designed from the ground up around the following three principles: |
BuoyOING has been designed from the ground up around the following three principles: |
520 |
\begin{quote} |
\begin{quote} |
521 |
\begin{enumerate} |
\begin{enumerate} |
522 |
\item the user should always see all link targets |
\item the user should always see all link targets |
996 |
% \\ |
% \\ |
997 |
% - RGB not perceptually uniform \\ |
% - RGB not perceptually uniform \\ |
998 |
% - Perceptually uniform color space: |
% - Perceptually uniform color space: |
999 |
% distance corresponds to perceived differece of color \\ |
% distance corresponds to perceived difference of color \\ |
1000 |
% - Usually color perception is divided to lightness, hue, and saturation \\ |
% - Usually color perception is divided to lightness, hue, and saturation \\ |
1001 |
% - CIELAB\cite{cie86colorimetry}\\ |
% - CIELAB\cite{cie86colorimetry}\\ |
1002 |
% |
% |
1229 |
|
|
1230 |
%In the following, we evaluate the implications and tradeoffs |
%In the following, we evaluate the implications and tradeoffs |
1231 |
%of different |
%of different |
1232 |
%aspects in the use of unique backround textures. |
%aspects in the use of unique background textures. |
1233 |
|
|
1234 |
\subsection{Overall appearance of the resulting textures} |
\subsection{Overall appearance of the resulting textures} |
1235 |
|
|
1257 |
but only the more modest goal of zooming within a range |
but only the more modest goal of zooming within a range |
1258 |
that would be reasonable for a single PDF document, |
that would be reasonable for a single PDF document, |
1259 |
i.e., approximately 20-fold |
i.e., approximately 20-fold |
1260 |
difference between mininum and maximum zoom. |
difference between minimum and maximum zoom. |
1261 |
|
|
1262 |
%The background textures should be be recognizable at different scales. |
%The background textures should be be recognizable at different scales. |
1263 |
Although texture perception is scale-independent to some extent, |
Although texture perception is scale-independent to some extent, |
1385 |
textures, we need to have an appropriate comparison point. |
textures, we need to have an appropriate comparison point. |
1386 |
Pictures of natural objects would not be appropriate, |
Pictures of natural objects would not be appropriate, |
1387 |
because they cannot be generated in infinite amounts from seed |
because they cannot be generated in infinite amounts from seed |
1388 |
numbers and they would esasily yield undesirable semantic associations. |
numbers and they would easily yield undesirable semantic associations. |
1389 |
Lacking a better example, we shall use plain solid color backgrounds |
Lacking a better example, we shall use plain solid color backgrounds |
1390 |
as a baseline |
as a baseline |
1391 |
even though the colors of even a small set of randomly chosen colors would |
even though the colors of even a small set of randomly chosen colors would |
1429 |
|
|
1430 |
Thus, we chose to measure the recognition of only 15 target textures |
Thus, we chose to measure the recognition of only 15 target textures |
1431 |
or colors in the experiment. |
or colors in the experiment. |
1432 |
Our hypthesis is that the texture backgrounds are more recognizable |
Our hypothesis is that the texture backgrounds are more recognizable |
1433 |
than the solid colors. |
than the solid colors. |
1434 |
|
|
1435 |
\subsubsection{Method} |
\subsubsection{Method} |
1443 |
chosen for both the texture and solid color conditions. |
chosen for both the texture and solid color conditions. |
1444 |
The distribution of the solid colors was the same that is used for |
The distribution of the solid colors was the same that is used for |
1445 |
the texture colors except that the highest |
the texture colors except that the highest |
1446 |
lighness tail was de-emphasized to increase the otherwise low |
lightness tail was de-emphasized to increase the otherwise low |
1447 |
discriminability of very light, unsaturated colors. |
discriminability of very light, unsaturated colors. |
1448 |
|
|
1449 |
\emph{Procedure.} |
\emph{Procedure.} |
1532 |
%Our example application --- indeed, the very motivation for this work --- |
%Our example application --- indeed, the very motivation for this work --- |
1533 |
%is the focus+context view |
%is the focus+context view |
1534 |
%of a Xanalogical structure\cite{ted-xanalogical-structure-needed} |
%of a Xanalogical structure\cite{ted-xanalogical-structure-needed} |
1535 |
%showin in Fig.~\ref{figxanalogicalexample}. For a given document, |
%shown in Fig.~\ref{figxanalogicalexample}. For a given document, |
1536 |
%the view shows the fragments of other documents that |
%the view shows the fragments of other documents that |
1537 |
%are linked to the current one. |
%are linked to the current one. |
1538 |
%Here, unique background textures help the user notice --- without thinking --- |
%Here, unique background textures help the user notice --- without thinking --- |
1542 |
|
|
1543 |
We have presented a perceptually designed hardware-accelerated |
We have presented a perceptually designed hardware-accelerated |
1544 |
algorithm for generating recognizably unique backgrounds. |
algorithm for generating recognizably unique backgrounds. |
1545 |
The motivating example, the BuoyING user interface demonstrates |
The motivating example, the BuoyOING user interface demonstrates |
1546 |
that the method is at its most useful |
that the method is at its most useful |
1547 |
when the same document can be reached through |
when the same document can be reached through |
1548 |
several ways and fragments of documents are seen. |
several ways and fragments of documents are seen. |
1549 |
|
|
1550 |
Of course, we cannot hope to match in quality a unique graphical |
Of course, we cannot hope to match in quality a unique graphical |
1551 |
appearance designed by a human designer; |
appearance designed by a human designer; |
1552 |
magazines and web sites have long used skilfully designed |
magazines and web sites have long used skillfully designed |
1553 |
graphical elements to make themselves recognizable. |
graphical elements to make themselves recognizable. |
1554 |
However, our algorithm is able to generate an unlimited |
However, our algorithm is able to generate an unlimited |
1555 |
amount of unique backgrounds cheaply, making it possible |
amount of unique backgrounds cheaply, making it possible |