Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Lossy Gray Scale Compression Using Biased Adaptive Delta Modulation with Arithmetic Coding

IP.com Disclosure Number: IPCOM000045365D
Original Publication Date: 1983-Mar-01
Included in the Prior Art Database: 2005-Feb-06
Document File: 3 page(s) / 42K

Publishing Venue

IBM

Related People

Anastassiou, D: AUTHOR

Abstract

A solution to the problem of applying the compression technique of binary adaptive arithmetic coding to lossy gray scale image compression is set forth below. Binary arithmetic coding is in (*). Arithmetic coding has until now been used only for lossless coding. the proposed scheme is based on Adaptive Delta Modulation with the additional feature that the quantization decisions are biased to achieve high compression. This technique can be used to enhance compression in all image coding applications in which Adaptive Delta Modulation is an acceptable coding method.

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 54% of the total text.

Page 1 of 3

Lossy Gray Scale Compression Using Biased Adaptive Delta Modulation with Arithmetic Coding

A solution to the problem of applying the compression technique of binary adaptive arithmetic coding to lossy gray scale image compression is set forth below. Binary arithmetic coding is in (*). Arithmetic coding has until now been used only for lossless coding. the proposed scheme is based on Adaptive Delta Modulation with the additional feature that the quantization decisions are biased to achieve high compression. This technique can be used to enhance compression in all image coding applications in which Adaptive Delta Modulation is an acceptable coding method.

As explained in detail in (*), arithmetic coding requires evaluation, for every bit to be coded, of the probabilities that it is 0 or 1 as relative proportions of precollected counts N(0)(S) and N(1)(S). These counts roughly specify the number of previous occurrences of a 0 or a 1 each time a bit-configuration (state) S around the pixel in question has occurred.

The bilevel image resulting from the 0's and 1's of the delta modulation can be further compressed using arithmetic coding to yield an encoding rate of less than 1 bit/pixel. It is proposed that in order to achieve a high compression, the following technique of biasing be used at the quantization stage of delta modulation.

First, an examination is performed for each pixel on whether a. the encoded area is smooth (determined by some indicator of

smoothness from the coded history), and

b. the quantization decision (i.e., whether the corresponding

bit of the delta modulation is 0 or 1, making the quantized

value smaller or greater than the predicted one) does not

lead to a significant difference in the resulting

quantization error.

If both the above conditions are satisfied, then the quantiza...