JPEG 2000

Dr. Daniel Lee
ISO SC29/WG1 (JPEG)
dlee@yahoo-inc.com

Editors’ Note
This interview is with Dr. Daniel Lee, who currently convenes ISO’s JPEG group. The ISO SC29/WG1 committee covers JPEG and JBIG. For those of you who are not familiar with the standard, you may want to visit the JPEG 2000 Web site for a brief overview of the content of the standard and links to related information. The JPEG 2000 Source site provides additional background information on who is involved in the development. We anticipate including items in future issues on the implications for implementing this standard and related topics.

Could you briefly describe JPEG 2000? Why should our readers pay attention to JPEG 2000? Who have the major players been in the development of JPEG 2000? What standards bodies are involved? 

JPEG 2000 is a new image coding standard developed by the International Standardization Organization (ISO) that serves the needs for imaging applications that the original JPEG standard does not meet. It is also an International Telecommunications Union (ITU) standard, so it has both an ISO and ITU number. It is hence designated as ISO/IEC 15444 and ITU-T T.800 series of standards.

JPEG 2000 is important because it is the cumulative work of over 100 imaging experts from 18 countries and represents state-of-the-art development in the imaging coding community. The players include leaders in the digital imaging industry, leading academic and industry research institutions, and government research institutions and agencies. 

How are JPEG 2000 software or vendor products certified as JPEG 2000 compliant?

Part 4 of the JPEG 2000 standard (ISO 15444-4) deals with conformance testing, where procedures for compliance to JPEG 2000 implementations are defined. 

How well suited is JPEG 2000 for the range of cultural objects (e.g., texts, manuscripts, photographs, art objects, etc.) being digitized by libraries, archives, and museums today? Is it, for example, a good choice for text and other image data that are characterized by edge detail? Is it a good alternative to Group 4 for 1-bit data? 

One of the design requirements for JPEG 2000 when it was first conceived was that the standard handle a wide variety of images, including those used to capture cultural objects like texts, manuscripts, photographs, art objects, etc. The technology adopted by JPEG 2000 renders itself to handle a wide range of image fields, including image data that are characterized by edge detail. It is not meant to be a replacement for Group 4 Facsimile standard, which is designed specifically for binary images (e.g., black-and-white texts). Even for those kinds of data, I would refer the readers to a newer and much more powerful standard, called the JBIG 2 standard, which is also developed under the same ISO standardization body that developed JPEG 2000.

JPEG 2000 offers many features not provided by the current JPEG standard, including support for multiple resolutions, tiling, and region-of-interest coding. Can you describe the advantages of these features and what kinds of cultural heritage images might benefit most from their use?

There are indeed many new features in JPEG 2000 that offer advantages over the current JPEG standard. In the context of cultural heritage images, support of multiple resolutions enables very effective image archive applications, so that a person viewing a particular spatial region of the image can extract (view) and drill down on the details of that region without spending computing and networking resources to decode the entire image.

Tiling support enables encoding of a very large image with extremely fine resolution that is more powerful than the current JPEG standard. Region-of-interest coding enables the optimal use of encoding and networking resources to preserve certain regions of image data with as much detail (resolution) as intended without assigning the same resource to the entire image.

How good is JPEG 2000's compression efficiency, and how well does it retain image quality with increasing compression? How do these features compare with JPEG and other common image file formats? Does JPEG 2000 offer both lossless and lossy compression modes? Does its compression work well with various bit-depth files? 

JPEG 2000 is designed to achieve excellent compression efficiency by using very advanced image coding techniques. It is particularly designed to retain the best image quality at each compression level. In technical terms, this is called rate-distortion optimization. Compared to current JPEG and other common image file formats, we have seen better compression efficiency, anywhere from 30% to 60%, bearing in mind that compression efficiency is image dependent.

JPEG 2000 offers not only lossless and lossy compression modes, it is uniquely designed such that an application can turn from a high-compression (lossy) mode down to lossless mode continuously.

JPEG 2000 works well at a wide range of image bit depths, as this is also one of the design features of the standard.

JPEG 2000 is still a work in progress. Some parts are incomplete, and those sections that have been completed have been subject to several technical corrections and amendments. Is JPEG 2000 ready for still-image users, who require stability and contemplate a long-term commitment to their images?

JPEG 2000 Part 1: Core System has been published for over a year now. Several minor technical corrections and amendments have since been published. It is absolutely ready for use. As an official ISO (jointly with ITU) standard, it is subject to a rigorous maintenance process to ensure that the standard is well maintained over its lifetime. The standard as designed is expected to serve image applications for at least the next 10 years.

A growing number of cultural heritage institutions have committed to the long-term maintenance of large image collections. Before migrating to another format, they will want assurance that the new format offers compelling functionality to merit the effort and expense of switching. A number of file formats, (e.g., MrSID, DjVu, LuraWave) offering wavelet compression and some of the same features as JPEG 2000 have appeared in recent years. Is there anything about JPEG 2000 that distinguishes it from its competitors and deserves special consideration as a replacement for these?

JPEG 2000 is based on the most-advanced image coding system. The technology that is incorporated into the standard has gone through rigorous testing to ensure that the performance goals are met. It has a rich set of features that can be considered tools that image applications can apply to provide the best solutions for a wide range of imaging needs. The standard has a set of compliance measures to ensure compatibility and interoperability. Hence it is the “tried and true” image standard. Indeed, many vendors are adopting JPEG 2000 and converting their proprietary formats to JPEG 2000.

Many cultural heritage institutions store their master image files in TIFF format, while using JPEGs to deliver the same images over the Web. Does JPEG 2000 have the potential to offer all these features in a single format and obviate the need to maintain two sets of files for each image? Should cultural heritage institutions consider using JPEG 2000 as a single preservation and delivery format?

JPEG 2000 offers a very comprehensive set of features in a single format and obviates the need to maintain two sets of files for each image. I definitely recommend that cultural heritage institutions use JPEG 2000 as a single preservation and delivery format.

What kinds of tools are available right now for the creation, manipulation, and delivery of JPEG 2000 images?

A variety of software tools—from many countries—are now available for creation, manipulation, and delivery of JPEG 2000 images. These tools include encoding, decoding, file conversions, and online image manipulation tools. A search in the World Wide Web will show where one can find these tools. 

Some examples:

http://www.leadtools.com/SDK/Raster/Raster-Addon-JPEG2000.htm

http://www.algovisionluratech.com/products/lurawave_jp2/browser/index.jsp?language=3

http://www.algovisionluratech.com/products/lurawave_jp2/photoshop/index.jsp?language=3

http://www.fnordware.com/j2k/

http://www.pegasusimaging.com/imagepressj2k.htm

http://www.elysium.ltd.uk/jpeg_2000.html

At the Image Compression Symposium, Steve Kerr and Bernie Brower [see resources listed at end] suggested that: “JPEG 2000 is more than a change in compression.  It is a paradigm shift in how we collect, store, transmit, and use image information.” How would you respond to that statement?

Yes, this is indeed a shift from the traditional image coding paradigm present when the original JPEG was developed. In a distributed world, where there is a high level of connectivity, it is a paradigm in which images and the resources to support the imaging applications are distributed. One no longer deals with sending and receiving images from just one source to another. Images can be coming from multiple sources, and the resources to manipulate, process, and perform applications are also coming from multiple sources. We need a new approach to handling image data so as to accommodate future applications in this new paradigm.

Could you describe ISO's overarching strategy for using scaleable approaches for archiving and delivering multimedia? How does JPEG 2000 fit into that strategy?

JPEG 2000 has been designed with scalability as the top requirement—multiple solution, spatial scalability, quality scalability (so-called signal-to-noise scalability), bit-stream scalability, file format and metadata scalability. JPEG 2000 is the building block for archiving and delivery of visual media in the overall ISO standardization activity.

JPEG 2000 offers numerous features not available in JPEG. Yet experience has shown that technological superiority is not necessarily sufficient when it comes to acceptance of new formats. For example, PNG has numerous advantages over GIF, yet it has failed to win large numbers of converts. Other potential formats, such as Flashpix, simply failed to catch on. What is the JPEG 2000 committee doing to ensure that JPEG 2000 will be more quickly and fully embraced by users and toolmakers?

The JPEG 2000 committee understands that  technology alone cannot make a standard pervasive. A standard needs the support of the application developers and the user community. To that end, the committee has published the Part 5, ISO 15444-5, which is the reference software implementation for JPEG 2000, so that developers can use it for applications development. The experts from the committee have been actively organizing conferences and  seminars and participating in major imaging conferences to promote the development of JPEG 2000. Industry groups such as the I3A have adopted JPEG 2000 as their standard of choice and promote it in trade events and exhibitions. Researchers from academic institutions have published books and articles on JPEG 2000 that reach out to the research community. All these would certainly help the pervasive adoption of JPEG 2000.

One preservation requirement is to preserve the integrity of the original file (object) over time. A number of institutions have invested in creating "libraries" of TIFF RGB masters. What would happen in a TIFF RGB to JPEG 2000 to TIFF RGB transformation? Would the output match the input? If not, what would potentially be lost?

Preservation of the image quality is one of the key features in the development of the JPEG 2000 standard. In the JPEG 2000 file format, very detailed considerations have been given to the correctness and the preservation of the Color Space representation/transformations of the image data. This information is encapsulated into the file format so that when the image data are passed among different applications (e.g., from scanning to printing), the fidelity is never lost.

At the moment, decoding a J2K, JP2, or JPX file with a Web browser requires a plug-in. Widespread use and acceptance of JPEG 2000 will probably require high-quality native support in the mainstream browsers. PNG has had native browser support for over five years, yet some of its best features are still not properly supported in popular browsers. Has Microsoft or AOL (Netscape) committed to providing full native support for any of the JPEG 2000 file formats? If so, when do you expect it to appear?

The JPEG 2000 committee has regular contacts with major browser manufacturers and will continue to seek their support in adding JPEG 2000 decoders into the browsers’ native functions.

The official JPEG 2000 Web site indicates that the core coding system is not patent free, though it is intended to be royalty and license-fee free. Should potential users be concerned that future assertion of patent rights may cause problems, such as happened with GIF images?

Throughout the development process of JPEG 2000, the committee has followed the ISO guidelines concerning technology selection. The committee has obtained the generous offer of royalty and license-fee free conditions by its technology contributors. While one can never be certain that there will never be any assertion of patent rights that may cause problems in the future, there should be some degree of confidence that the standard as developed by so many experts from so many countries should be safe for users to adopt.

TIFF is widely used by cultural heritage institutions for master image files. The main TIFF specification hasn't been updated in over a decade. This may a blessing, as regular updates often introduce "feature creep" with resulting incompatibilities. On the other hand, a dead image format will eventually suffer obsolescence. Should users be resigned to the need for periodic migration to new formats? What are the signs that an image format is reaching the end of its life expectancy?

The clear sign that an image format is reaching the end of its life expectancy is that it no longer supports new applications, particularly those that come out from new paradigms of application environments.

Suggested Resources

“Library Potential Impacts,” by Steve Kerr and Bernie Brower, Image Compression Symposium.

“ISO JPEG 2000 Standards Efforts,” by Gordon Ferrari, Image Compression Symposium.

“NITFS JPEG 2000 Implementation Schedule/Events,” Bandwidth Compression Symposium.

“The Next Generation of Compression JPEG 2000,” by Bernie Brower, Image Compression Symposium.

Publishing Information

RLG DigiNews (ISSN 1093-5371) is a newsletter conceived by the members of the Research Libraries Group's PRESERV community. Funded in part by the Council on Library and Information Resources (CLIR) 1998-2000, it is available internationally via the RLG PRESERV Web site. It will be published six times in 2002. Materials contained in RLG DigiNews are subject to copyright and other proprietary rights. Permission is hereby given for the material in RLG DigiNews to be used for research purposes or private study. RLG asks that you observe the following conditions: Please cite the individual author and RLG DigiNews (please cite URL of the article) when using the material; please contact Jennifer Hartzell, RLG Corporate Communications, when citing RLG DigiNews.

Any use other than for research or private study of these materials requires prior written authorization from RLG, Inc. and/or the author of the article.

RLG DigiNews is produced for the Research Libraries Group, Inc. (RLG) by the staff of the Department of Preservation and Conservation, Cornell University Library. Co-Editors, Anne R. Kenney and Nancy Y. McGovern; Production Editors, Martha Crowe and Barbara Berger Eden; Associate Editor, Robin Dale (RLG); Technical Researchers, Richard Entlich and Peter Botticelli; Technical Coordinator, Carla DeMello; Technical Assistant, Kimberly Gazzo.

All links in this issue were confirmed accurate as of December 11, 2002.

Please send your comments and questions to RLG DigiNews Editorial Staff.