1、CCITT RECflN*T.AO 92 m 4862593 0579255 98T INTERNATIONAL TELECOMMUNICATION UNION CCITT THE INTERNATIONAL TELEGRAPH AND TELEPHONE CONSULTATIVE COMMITTEE TERMINAL EQUIPMENT AND PROTOCOLS FOR TELEMATIC SERVICES T.80 (09/92) COMMON COMPONENTS FOR IMAGE BASIC PRINCIPLES COMPRESSION AND COMMUNICATION - Re
2、commendation T.80 CCITT RECNN*T.80 92 W 4862591 0579256 816 FOREWORD The CCIT (the International Telegraph and Telephone Consultative Committee) is a permanent organ of the International Telecommunication Union (ITU). CCITT is responsible for studying technical, operating and miff questions and issu
3、ing Recommendations on them with a view to standardizing telecommunications on a worldwide basis. The Plenary Assembly of CCIT which meets every four years, establishes the topics for study and approves Recommendations prepared by its Study Groups. The approval of Recommendations by the members of C
4、CITT between Plenary Assemblies is covered by the procedure laid down in CCITT Resolution No. 2 (Melbourne, 1988). Recommendation T.80 was prepared by Study Group VI11 and was approved under the Resolution No. 2 procedure on the 18th of September 1992. CCIT NOTE In this Recommendation, the expressio
5、n “Administration” is used for conciseness to indicate both a telecommunication administration and a recognized private operating agency. O ITU 1993 Ail rights reserved. No part of this publication may be reproduced or utilized in any form or by any means, electronic or mechanical, including photoco
6、pying and microfilm, without permission in writing from the ITU. CCITT RECflNtT.80 92 48b259L 0579257 752 INTRODUCTION This Recommendation was prepared by CCITT SG VIII. The requirements, contained in clause 5, however, were forwarded to an informal CCIT-ISO/IEC “Joint Photographic Experts Group” (P
7、EG). The task of this Group was to develop a universal progressive/sequential compression algorithm for any type of colour model. During the development phase extensive testing and comparison of various techniques showed, that although the original goal was technically feasible, the initiai solution
8、 did not produce the best compression and image-quality results for all1 the image types addressed. Therefore it was decided that PEG should develop a progressive/sequential continuous-tone color and gray scale image compression technique (described in Recommendation T.8 1), while a newly formed sep
9、arate group, CCIT-ISO/IEC “Joint Bi-level Image Group” (BIG) should develop a separate (but from a functionality point of view similar) progressive/sequential bi-level (limited multi-level) image-compression technique (described in Recommendation T.82). Thus, there are many commonalities between the
10、 two techniques (such as their functionality, the use of a joint entropy coder, etc.), but the two techniques are different from an aigorithm point of view. Reference to the PEG and JBIG developed algorithm are provided in clause 2, Normative references. Both techniques were jointly standardized by
11、CCITT and ISO/IEC. Recommendation T.80 (09/92) i CCITT RECMNIKT.80 92 4862593 0579258 699 W Recommendation T.80 COMMON COMPONENTS FOR IMAGE COMPRESSION AND COMMUNICATION - BASIC PRINCIPLES (1992) The CCITT, considering (a) that the CCITT seeks the most efficient methods of coding, communication and
12、storage of images; (b) that there is a growing interest in telematic services (such as facsimile, videotex, teleconferencing) and other CCITT applications to transmit images on telecommunication networks with varying speeds; (c) that there is a growing demand for multilevel (including gray-scale and
13、 continuous tone color) imagery; (d) that there is a trend to multifunctional terminals participating in more than one telematic services and/or other CCIT and/or data-processing applications; (e) that with the introduction of new telecommunication network technologies, such as ISDN, there is a grow
14、ing tendency to provide multiple CCIT services and/or other applications on the same telecommunication network; (f) that there is an urgent demand for interworking among telematic services (and other CCIT applications) used for image-communication and/or image-storage; (8) that there is a strong con
15、vergence of image-processing, image-storage and image-communication techniques; (h) that there is strong tendency toward introducing “soft-copy” (i.e. screen oriented) image-communication servicedapplications; (i) that there is a need for harmonization of “soft-copy” (i.e. screen oriented) and “hard
16、-copy” (Le. print- oriented) image presentation modes, within the same, or among several CCIT services/applications, recognizing (i) that while several established techniques (see Recommendations T.4 and T.6) for efficient coding of images containing black and white text and line-drawings exist in C
17、CITT without using new image-compression techniques, the needs expressed above can only partially be satisfied, unless new image-compression techniques are provided, unanimously recommends that in addition to image-compression techniques defined in Recommendations T.4 and T.6 a new generation of ima
18、ge-compression techniques containing new or improved properties (including, but not limited to: gray- scde/colour, progressive/sequential image build-up, lossyllossless compression, be defined according to the T.80-Series Recommendations. 1 Scope The T.80-Series Recommendations are applicable to a w
19、ide range of CCIT applications which require use of compressed still images. Such images can range from bi-level to continuous-tone single-color and multi-color still digital image. It is applicable to a wide range of CCIT applications which require use of compressed images. This Recommendation is t
20、he introduction to the T.80-Series Recommendations. Recommendations T.81 and T.82 describe the individual image compression techniques. Recommendation T.80 (09/92) 1 CCITT RECNN*T-80 92 4862593 0579259 525 This Recommendation describes the basic principles of the requirements to be satisfied by the
21、new compression techniques, the relation of those compression techniques to the individual CCITT applications and services. This Recommendation does not specify the individual compression techniques to be used. Neither does it specify, for a given application, a complete representation of coded imag
22、e data, nor specify an application that is using any of the new compression techniques. 2 Normative references The following CCITT Recommendations and IO/IEC International Standards contain provisions which, through reference in this text, constitute provisions of this Recommendation. At the time of
23、 publication, the editions indicated were valid. All Recommendations and International Standards are subject to revision. i CCITT Recommendation T.81 (1992) I ISO/IEC 10918-1:1992, Information technology - Digital compression and coding of continuous-tone still images - Part I: Requirements and guid
24、elines. 2 CCITT Recommendation T.82 (1992) I ISO/IEC 11544:1992, Znfonnation technology - Coded Representation of Picture and Audio Information - Progressive Bi-level Image Compression. Informative references to other related CCITT Recommendations, ISO/IEC Standards and application oriented CCITT Re
25、commendations are provided in the References clause. 3 Definitions For the purposes of this Recommendation, the following definitions apply. 3.1 bi-level image An image with two distinct levels of predefined colors (e.g. a typical facsimile image is a bi-level black-and- white image). 3.2 bit-plane
26、An array (or “plane”) of bi-level symbols constructed from an image by choosing a particular bit from each pixel. 3.3 byte Eight bits of data. 3.4 coding process A general term referring to an encoding process, a decoding process, or both. 3.5 color image A bi-level or multi-level image containing m
27、ore than one color component, defining the full color specification of that image. 3.6 color component _- One of the two-dimensional arrays which comprise a color image. The meaning of the color component is defined by the respective color model used. 2 Recommendation T.80 (09/92) CCITT RECMN*T-BO 9
28、2 4862591 0579260 247 3.7 color model A defined way of describing colors. The color models selects the appropriate color components that are used to describe a specific color of an image sample. 3.8 complete representation of a coded image A T.81 or T.82 data-stream embedded in the application (data
29、) environment of a given CCITT application or service (e.g. in Photographic Data Syntax of Videotex). 3.9 compression Reduction in the number of bits used to represent source image data. 3.10 continuous-tone image An image whose color components contain a sufficient number of bits (color steps) per
30、pixel, that the subjective appearance of the image appears continuous (e.g. 256 levels of gray provide the subjective impression of a true gray-scale image). 3.11 decoder An embodiment of a decoding process. 3.12 decoding process A process which takes as its input compressed image data and outputs e
31、ither an equal to or similar to the original uncompressed image. 3.13 (digital) (still) image A two-dimensionai spatial array of image component data without time dimension, or a set of such arrays. 3.14 dithering A technique applied to single bit-plane images, which makes the single bit-plane image
32、 appear subjectively as a continuous-tone image. When dithering is used in conjunction with a fixed image resolution, the result is a perceived “increase” in gray-levels. At the same time, however, the perceived image resolution decreases. 3.15 encoder An embodiment of an encoding process. 3.16 enco
33、ding process A process which takes as its input an original image as its input and outputs compressed image data. 3.17 entropy coder Any lossless method for compressing data. 3.18 fixed-resolution progressive mode A progressive image encoding/decoding process, whereby the quality of the image is imp
34、roved by each successive step, while the original resolution of the image remains unchanged. 3.19 gray-scale image A continuous-tone image with one color component only. Recommendation T.80 (09/92) 3 CCITT RECMN*T.BO 72 4Bb2573 0579263 383 3.20 half-toning An artificial coloring process (similar to
35、dithering) used in printing that is applied to the bit-planes of the primary printing color components, and which results in a printout that subjectively appears to have continuous-tone color characteristics. 3.21 “hard-copy99 image An image which has been printed on an output device (a fixed, highe
36、r resolution image is usually expected). 3.22 hierarchical progressive mode A progressive encoding/decoding process, such that, when decoding, a lower resolution image is decoded first, then by successive decoding steps, the resolution of the image is increased horizontally and vertically by predefi
37、ned factors (usually by 2). in Recommendation T.82 this is the only type of “progressive” mode defined. 3.23 image data Either source image data or reconstructed image data. 3.24 interleave A sampling of different color components with different sampling frequencies. It results in different size of
38、image resolutions for different color components. If correctly applied, remarkable compression gain can be achieved. 3.25 limited multi-level image A term denoting that the precision with which each color component of a color image, or the single component a gray-scale image is specified, is limited
39、 to about two to seven bits. 3.26 lossless A descriptive term for encoding and decoding processes in which the output of the decoding procedure is (data) identical to the input data to the encoding process. 3.27 lossy A descriptive term for encoding and decoding processes which are not lossless, i.e
40、. the decoding process output (data) is not identical with to the encoding process input (data). However, a lossy encoding and decoding process may produce an image with subjectively excelient image quality. Lossy coding compression is often applied in order to gain additional image compression. 3.2
41、8 multi-level image A term for describing the precision-range of color components of a gray-scale or color image. Multi-level includes precision starting from about three levels (bits) to usually eight levels, but for some special applications 12, 16 or even higher. 3.29 picture A bounded representa
42、tion of visual information. 3.30 picture coding The presentation in digital form of a picture, particularly for storage, interchange or processing. 4 Recommendation T.80 (09/92) CCITT RECMN*T.BO 92 M 4862591 O579262 OLT 3.31 pixel One picture element of an image which is described by a rectangular a
43、rray of such elements. 3.32 pixel aspect ratio Definition of the actual shape of a single pixel, as defined by an application using the T.80-Series Recommendations. 3.33 photographic coding A technique for representing pictures by means of pixels, using high resolution and a wide range of grey level
44、s andor colors giving a natural appearance. 3.34 progressive behaviour A coding technique shows progressive behaviour if an image is first coded as a lowest resolution layer image and then is successively increased in resolution by means of differential layer images. 3.35 progressive coding A method
45、 of describing the entire image is first described at a lower image quality (i.e. lower resolution image or an image without details) and then step-by-step successively build up and improve to the desired final image quality (Le. final image resolution or image with many details). 3.36 progressive-c
46、ompatible sequential coding In Recommendation T.82, a method of coding an image, in which the image may be segmented into stripes, the image stripes are coded in sequence, and within each sequence, and within each stripe the image is coded to full resolution progressively. This is compatible by stri
47、peAayer data reordering with progressive coding. 3.37 quantization A process to represent an analogue signai source by an appropriate digital value. 3.38 sample A sample represents one element in the two-dimensional array which comprises a component. In the image conversion process of the analogue s
48、ource image into a digital image, “samples” are taken, line by line, in regular intervals. Depending on the application-dependent sampling process, the relation of samples to pixels is determined by factors, such as color model, interleave, sampling geometry, etc. 3.39 sequence behaviour A coding te
49、chnique shows Sequential behaviour if portions of the image near the top are completely described before portions below have been described at ail. 3.40 sequential coding A method of describing an image in which portions of the image near to the top are completely described before portions below have been described at all. 3.41 single-progression sequential coding In T.82, a method of coding an image, such that the image is fully coded in a single resolution layer, line by line, from left to right and top to bottom, without reference to any lower resolution images. This is com
