Browse Prior Art Database

Lossless Color Image Compression Technique for Multimedia Applications

IP.com Disclosure Number: IPCOM000113994D
Original Publication Date: 1994-Oct-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 4 page(s) / 158K

Publishing Venue

IBM

Related People

Chen, T: AUTHOR [+2]

Abstract

In multimedia applications, quite often color images need to be stored away for later retrieval and displayed on a screen. In addition, transmitting these images from one location to another remote location over a network is becoming very common.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 40% of the total text.

Lossless Color Image Compression Technique for Multimedia Applications

      In multimedia applications, quite often color images need to be
stored away for later retrieval and displayed on a screen.  In
addition, transmitting these images from one location to another
remote location over a network is becoming very common.

      A high resolution full color image could have in excess of one
million pixels which translates into over 24M bits of data.  To store
such images in an uncompressed form would require a substantial
amount of storage space.  One approach to effectively utilize the
storage space is to compress the information data before storing.
Also, transmitting a compressed color image requires less bandwidth
than if the image was not compressed.  Unlike lossy compression
techniques which lose original image information after decompression,
this disclosure devises a lossless compression scheme which achieves
a high level of data compression as well as preserving all original
image information.  Lossless color image compression is essential for
a number of applications, for example it is especially important to
the medical community where resolution and precise details are
crucial for medical diagnosis.

      The compression scheme disclosed here takes the advantage of
the fact that adjacent red, green and blue contents of a pixel do not
differ drastically.  By encoding the color difference of adjacent
pixels, the amount of image data can be reduced significantly without
losing any original image information at all.  Based on this encoding
principle, a lossless compression algorithm is developed to
effectively compress color image data for applications that require
full retention of the original image data.

      If one examines a color image closely one could conclude that
the primary color content of most adjacent pixels do not differ
significantly.  By taking advantage of this characteristic of a color
image, one can encode the color content difference of adjacent pixels
in 0, 2, or 4 bits without losing any information, hence preserving
the original image.  In addition, a better utilization of storage
space is attained.  The resulting compression ratio is dependent on
the number of encoding bit (0, 2 or 4) chosen for the algorithm and
the characteristics of the image.  The discussion below details this
lossless compression technique.

      In a full color graphics raster-scan system, the Red (R), Green
(G) and Blue (B) contents of each pixel in an image are stored in the
frame buffer as three contiguous bytes (24 bits) as depicted in Fig.
1.

      The pixel arrangement and pixel addressing scheme in the frame
buffer depend widely on the type of memory chosen and on the
particular design.  Fig. 2 is an example of a pixel arrangement used
here solely to demonstrate the compression scheme under discussion.
The compression technique discussed here is independent of the design
of the frame buff...