Browse Prior Art Database

Image Expander and Compressor

IP.com Disclosure Number: IPCOM000038702D
Original Publication Date: 1987-Feb-01
Included in the Prior Art Database: 2005-Jan-31
Document File: 4 page(s) / 29K

Publishing Venue

IBM

Related People

Liu, CN: AUTHOR [+2]

Abstract

Image data generally originates at a camera or an array of sensors. Image data can also originate from computer simulations. The demand for storage, processing and transmission of image data is increasing very rapidly. High performance and/or special processors are built to process image data, but the high data rates of many image applications create requirements that tend to exceed available system capacities. In order to minimize the storage and bandwidth problems in connection with the storage and transmission of image data, image compression has been investigated and a number of algorithms have been implemented.

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 35% of the total text.

Page 1 of 4

Image Expander and Compressor

Image data generally originates at a camera or an array of sensors. Image data can also originate from computer simulations. The demand for storage, processing and transmission of image data is increasing very rapidly. High performance and/or special processors are built to process image data, but the high data rates of many image applications create requirements that tend to exceed available system capacities. In order to minimize the storage and bandwidth problems in connection with the storage and transmission of image data, image compression has been investigated and a number of algorithms have been implemented. The algorithms that have been developed require a low level of complexity in their implementations, the achievement of a high compression ratio and the minimization of distortions/artifacts in the compressed- decompressed images. On the other hand, many image applications also require that originally small images be expanded (e.g., zoomed) for interactive interpretation. This article is concerned with a new solution to both problems. A new method to compress image data with a high compression ratio is presented. It is easily implementable, produces a minimal amount of distortion, and contains a number of parameters which can be used to produce desirable results for different applications. The reconstruction method described in this article can be used to expand any small image to a desired size for interpretation. A principal consideration in employing data compression is the effect on image fidelity. In spite of considerable progress in data compression research, a satisfactory criterion for subjective image quality is still lacking. The signal-to-noise ratio is frequently used, and we will also use it to compare the reconstructed image with the original image. The underlying philosophy of our approach is to approximate local structure of the image by a two-dimensional polynomial of degree kx in the x direction and ky in the y direction. The choice of the coefficients of this polynomial is such that square error is minimized in a certain neighborhood Sij of every interpolated point (i,j). This basic approach can be applied for two separate purposes:
1) Estimating an image by interpolation based on an

arbitrary number of neighboring points. 2) Compressing the image by sampling (or other linear transformation) and then using the proposed

interpolation techniques for the image reconstruction.

This approach to the image compression can be viewed as

the combination of prediction and transform techniques

which are usually used as alternative to each other. A general solution to the above-mentioned problems will be described followed by a much simplified solution for practical applications. For the sake of simplicity, assume that kx = ky = k and Sij is a square window of the size mxm, m>k. We want to interpolate the image at the point (i,j) based on N>k2 _ samples in the window Sij:

1

Page 2 of...