Browse Prior Art Database

Color Graphics Picture Segmentation

IP.com Disclosure Number: IPCOM000036018D
Original Publication Date: 1989-Aug-01
Included in the Prior Art Database: 2005-Jan-28
Document File: 4 page(s) / 35K

Publishing Venue

IBM

Related People

Anastassiou, D: AUTHOR [+3]

Abstract

Compressibility of a color graphics image, such as a color presentation foil, is enhanced by an analytical procedure which segments the image into N regions each characterized by one of N different colors that occur most frequently in the picture elements (pixels) of the original image. The number N can vary from 2 to a maximum value considered to be a practical upper limit for image compression purposes. The specific value of N for any given image depends upon the variety and distribution of colors in the original image and is determined in the course of executing the color analyzing procedure.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 47% of the total text.

Page 1 of 4

Color Graphics Picture Segmentation

Compressibility of a color graphics image, such as a color presentation foil, is enhanced by an analytical procedure which segments the image into N regions each characterized by one of N different colors that occur most frequently in the picture elements (pixels) of the original image. The number N can vary from 2 to a maximum value considered to be a practical upper limit for image compression purposes. The specific value of N for any given image depends upon the variety and distribution of colors in the original image and is determined in the course of executing the color analyzing procedure.

As an example, it will be assumed that the colors of all the pixels in the original multicolor image are formed of three basic color components: red(R), green(G) and blue(B). Hence, the particular color of each individual pixel in the image can be defined as a combination of R, G and B coordinate values. The color component value of each pixel can be expressed as an eight-bit code. However, to avoid an unduly complex implementation, only the two highest-order bits of each color component code are considered in this process. This means that the 256 possible values of each color component will be considered not individually but as members of four possible groups,

(Image Omitted)

as follows:

Group Number Range of Values

0 0 to 63

1 64 to 127

2 128 to 191

3 192 to 255

Henceforth, in this description these group numbers will be treated as coordinate values, each representing a typical color component value of its respective group.

Since each of the three color components can have any one of these four representatives, the total number of allowable combinations of color component values is 64. Symbolically this can be visualized as a cubical grid (Fig. 1), the three dimensions of which correspond to the three color coordinates R, G and B, respectively. Each of these color coordinates has four possible values, making 64 grid points within and on the surfaces of the cube. Each of these grid points will be referred to as a "color", and it will be understood that this term means a particular combination of allowable color component values. Experience has shown that 64 colors form a sufficiently rich set for the approximate, but acceptable, rendition of certain simple color foil pictures.

In general, it is not necessary to represent all 64 of the allowable colors in a compressed color image. Usually the distribution of colors among the pixels will tend to emphasize a relatively small number of dominant colors out of the maximum number of 64, and for image compression purposes this small number

1

Page 2 of 4

of colors, N, can be used exclusively to form the output image which is to be compressed. The procedure described below explains how the original image can be segmented into N regions each characterized by one of the N dominant colors.

A 64-point three-dimensional (3-D) histogram, corresponding to the 3-D grid s...