Browse Prior Art Database

Dual Resolution Adaptive DPCM

IP.com Disclosure Number: IPCOM000050032D
Original Publication Date: 1982-Aug-01
Included in the Prior Art Database: 2005-Feb-09
Document File: 3 page(s) / 48K

Publishing Venue

IBM

Related People

Anastassiou, D: AUTHOR [+2]

Abstract

A high compression differential pulse code modulation (DPCM)-like coding technique for lossy gray-scale image compression is described below. A variable number of bits along with adaptive quantization are used to code each pixel (picture element) according to the smoothness of the local area of the pixel. The smoothness is determined by a function of the already coded history. This function is referred to as the "smoothness indicator". The scheme achieves an encoding rate of less than 1 bit/pixel by occasionally shifting to a "low resolution" coding mode. It operates best when a very high compression of slightly oversampled images is desired.

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 53% of the total text.

Page 1 of 3

Dual Resolution Adaptive DPCM

A high compression differential pulse code modulation (DPCM)-like coding technique for lossy gray-scale image compression is described below. A variable number of bits along with adaptive quantization are used to code each pixel (picture element) according to the smoothness of the local area of the pixel. The smoothness is determined by a function of the already coded history. This function is referred to as the "smoothness indicator". The scheme achieves an encoding rate of less than 1 bit/pixel by occasionally shifting to a "low resolution" coding mode. It operates best when a very high compression of slightly oversampled images is desired.

This DPCM-like scheme adaptively and automatically selects one of several modes of operation which result in the transmission of zero, one, two or more bits to encode a pixel. The shift is automatic because it does not require any marker bits but, rather, depends on information from the previously coded pixels. The different modes are used to give good compression while maintaining the local degradation of the coded image close to the threshold of distortion detectable by the visual system.

The smoothness indicator determines whether one or two bits will be used to code a pixel and also sets the actual (two or four) quantization levels used, which are arranged symmetrically around a predicted intensity value. Under certain conditions the low resolution mode is entered, in which case the encoder uses one bit to code a whole block of four pixels. On the other hand, sharp or unexpected edges will not be well represented by intensity values close to the predicted one. This has been a major problem of delta modulation for image cod...