588 |
|
|
589 |
%XXX: shorten by one half column |
%XXX: shorten by one half column |
590 |
|
|
591 |
In this section, we discuss the principles underlying |
In this section, we discuss the general principles |
592 |
|
we have used to derive |
593 |
our algorithm to generate unique |
our algorithm to generate unique |
594 |
background textures. |
background textures. |
595 |
To be useful, the unique backgrounds should be easily |
The goals are that the unique backgrounds should be easily |
596 |
distinguishable and recognizable, and should not |
distinguishable and recognizable, and that they should not |
597 |
significantly impair the reading of black text on top of it. |
significantly impair the reading of black text drawn on top. |
598 |
|
|
599 |
|
Our approach is an extension of |
600 |
|
the type of inversion approach that been used |
601 |
|
by Ware and Knight\cite{ware95texture} for inverting |
602 |
|
the earliest stage of the visual system%, |
603 |
|
%the spatial frequency detectors, |
604 |
|
in order to place |
605 |
|
a particular vector or scalar field of data |
606 |
|
in the texture ``channel''. |
607 |
|
|
608 |
The ability to distinguish a particular texture from a large set |
The ability to distinguish a particular texture from a large set |
609 |
depends on the distribution of textures in the set. |
depends on the distribution of textures in the set. |
746 |
%For example, pixels on a small area should correlate enough to |
%For example, pixels on a small area should correlate enough to |
747 |
%facilitate perception of contours. |
%facilitate perception of contours. |
748 |
|
|
|
A type of inversion approach has been used |
|
|
by Ware and Knight\cite{ware95texture}, for inverting |
|
|
the earliest stage of the visual system, |
|
|
the spatial frequency detectors, in order to place |
|
|
a particular vector or scalar field of data |
|
|
in the texture ``channel''. |
|
|
|
|
749 |
% To understand why it is possible to learn to discriminate particular |
% To understand why it is possible to learn to discriminate particular |
750 |
% backgrounds easily, consider the task of learning {\em one} background texture. |
% backgrounds easily, consider the task of learning {\em one} background texture. |
751 |
% This is a two-class problem. |
% This is a two-class problem. |