Browse Prior Art Database

Image Size Reduction Method using Frequency Information

IP.com Disclosure Number: IPCOM000114310D
Original Publication Date: 1994-Dec-01
Included in the Prior Art Database: 2005-Mar-28
Document File: 2 page(s) / 33K

Publishing Venue

IBM

Related People

Nakayama, Y: AUTHOR

Abstract

Disclosed is a technique whereby the size of a bitmap graphics image can be reduced without causing either thin dots or lines to vanish. Information of the frequency of colors is used to save pixels with infrequent color values that are surrounded by pixels with frequent color values, such as those forming a background.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 89% of the total text.

Image Size Reduction Method using Frequency Information

      Disclosed is a technique whereby the size of a bitmap graphics
image can be reduced without causing either thin dots or lines to
vanish.  Information of the frequency of colors is used to save
pixels with infrequent color values that are surrounded by pixels
with frequent color values, such as those forming a background.

      In the process of reducing the size of a bitmap graphics image,
detailed parts of the image tend to be lost, because pixels with
infrequent color values in the background have low priority in a
simple sampling algorithm.

      To solve this problem, the number of pixels of each color in
the entire image is counted before the reduction process.  A
Frequency table showing F(c) is provided for the next reduction
process, where F(c) is the number of pixels that have a color with
the value c.  A suitable fixed value L % (0 le L le 1) is provided as
a threshold value, to separate the background color from others.  In
the reduction process, the original image is divided into small
areas.  A sample pixel is selected from each area and <F(c sub 1 )>
over T is compared with L, where c sub 1 is the color value of the
sample pixel, and T is the total number of pixels in the image.

      If <F(c sub 1 )> over T is less than L, the area is compressed
into a single pixel that has the color value c sub 1.  If <F(c sub
1)>
over T is greater than or equal to L, the color values of th...