Browse Prior Art Database

Dot Modelling for Printing Grey Scale Images

IP.com Disclosure Number: IPCOM000120653D
Original Publication Date: 1991-May-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 3 page(s) / 136K

Publishing Venue

IBM

Related People

Todd, SJ: AUTHOR

Abstract

Disclosed is a model of a simple compensating dot-overlap which can be used during the pseudo half-toning error diffusion process to prevent patterning problems on a printer. It relates to the printing of grey- scale images on bilevel printers also to color bilevel per color printers and to photocomposers. It is applicable whether the basic pseudo half-toning technique is dithering white or blue noise, error diffusion peano, extended peano, Floyd Steinberg, or Floyd Steinberg with random weights.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Dot Modelling for Printing Grey Scale Images

      Disclosed is a model of a simple compensating dot-overlap
which can be used during the pseudo half-toning error diffusion
process to prevent patterning problems on a printer.  It relates to
the printing of grey- scale images on bilevel printers also to color
bilevel per color printers and to photocomposers.  It is applicable
whether the basic pseudo half-toning technique is dithering white or
blue noise, error diffusion peano, extended peano, Floyd Steinberg,
or Floyd Steinberg with random weights.

      Dots laid down by laser and other printers overlap giving rise
to two problems.  The blackness of the region is not linearly related
to the proportion of black dots, and with a fixed proportion of black
dots, the blackness depends on the exact pattern of dots.  If 50% of
pixels are black and they lie in a chessboard pattern, many printed
will give pure black output.  If they lie in stripes two pixels wide,
the same printer may give less that 70% blackness.  Standard
techniques conventionally used to solve the problem have
disadvantages.

                            (Image Omitted)

      Superpixels; the collections of black dots in groups which give
a strong repeating screen on the image, and Gamma correction that
fixes the first problem, to give overall correct brightness levels
but does not solve different patternings of dots.  Over a large area
of equal intensities with fixed dot proportion, the dot pattern may
change, creating a low frequency area noise.

      The new method is based on a pixel contamination model. Given a
pixel(i,j), we call area(i,j) the square area on the paper that the
pixel covers, and dot(i,j) the area that is made black by the printer
when pixel(i,j) = 1.  For simplicity assume that the area of each
pixel is 1, AREA(area(i,j)) = 1, where AREA measures area.

      If pixel(i,j) = 1, all of area(i,j) that belongs to pixel(i,j)
is assumed to be set black.  Also, a certain proportion of each of
the neighbouring pixels (i-1,j), (i+1,j), (i,j-1), (i,j+1) is set
black by the printed dot(i,j) overlapping area(i,j).  The
contamination factor x is the proportion of the total area of each of
these neighbouring pixels that is covered by overflow from dot(i,j),
e.g.,
      x = AREA( pixel{(i,j) intersect area(i,j+1) )

      In the model it is assumed that if a pixel is contaminated by
several neighbours, the total contamination is the sum of the
independent contaminations.  Also assumed was that there is no
contamination of diagonal neighbours. Equivalently assumed is that
dot (i,j) does now overlap with dot*i,j+2), dot (i+1,j+1), etc.  This
simplification could be lifted by a minor complication to the
algorithm.  For example, a 256-entry look-up table could give total
contamin...