G U E S T E D I T O R I A L Special Section on Internet Imaging Giordano Beretta Hewlett-Packard Company 1501 Page Mill Road, 4U-6, Palo Alto, CA 94304 USA E-mail: beretta@hpl.hp.com Raimondo Schettini ITIM CNR Via Ampere 56 I-20131 Milano, Italy E-mail: centaura@itim.mi.cnr.it Internet imaging differs from other forms of electronic imaging in that it employs an internet (network of net- works) as a transmission vehicle. How- ever, the internet is only one compo- nent (albeit a major one) in the total imaging system. The total system com- prises client applications internet- worked with server applications, as well as offline authoring tools. The internet is an evolving commu- nication system. Its functionality, reli- ability, scaling properties, and perfor- mance limits are largely unknown. The transmission of images over the inter- net pushes the engineering envelope more than most applications. Conse- quently, the issues we are interested in exploring pertain to all aspects of the total system, not just images or imag- ing algorithms. This emphasis on systems is what sets internet imaging apart from other electronic imaging fields. For a local imaging application, even when it is split between a client and a server linked by an Ethernet, a system can be designed by stringing algorithms in a pipeline. If performance is an issue, it is easy to identify the weak link and re- place it with a better performing com- ponent. On the internet, the servers are un- known, the clients are unknown, and the network is unknown. The system is not easily predictable and the result is that the most common problem today is scalability. To be successful one has to follow a top-down design strategy, where the first step is a detailed analy- sis of the problems to be solved. When a solution is invented, algorithms are selected to produce a balanced sys- tem, instead of choosing algorithms of best absolute performance as is done in bottom-up approaches. The paper on the Visible Human by Figuiredo and Hersch is a good ex- ample illustrating these fundamentals. Today, storing a 49-Gbyte 3-dimen- sional volume is not hard, and a RAID disk array can deliver fast access times. However, storage space and seek time are not the limiting factors for the extraction of ruled surfaces from large 3-dimensional medical images. The problem is one of load balancing, which requires detailed performance measurements for scalability. Eventu- ally, a specialized parallel file striping system must be designed and opti- mized. Implementing and maintaining a system that must grow as more data becomes available and as surgeons re- quire new staging techniques for tu- mors is practical only in a centralized solution served on the internet. After e-mail, the most popular appli- cation on the internet is the World Wide Web, which is a hypertext system and as such is useful only when it can eas- ily be navigated through a visual interface,1 and search results are pre- sented in a context,2 as is illustrated for example by the KartOO search engine. Navigation requires structure,3 and al- though techniques such as ontologies have been known for years, the par- ticularities for decoupling and splicing ontologies are not yet sufficiently understood.4 In a recent paper, the Jörgensens have described the challenges of de- veloping an image indexing vocabulary,5 and yet we know that tax- onomies are not sufficiently powerful for efficiently finding related information through navigation.6 Progress in bio- informatics has given us new computa- tional tools that will allow the develop- ment of new collaborative structuring methodologies based on ontologies. Another example of how wrong things can go when the fundamentals of internet imaging are not understood is content-based image retrieval (CBIR) systems. Today they are part of all the major search engines on the in- ternet, and anyone who has tried to use them for real work has experienced how useless they are. Although over the years a number of CBIR algorithms has been proposed, none has stood out as being particu- larly robust, despite the fact that each claims to perform best on some bench- mark. Unfortunately there is no univer- sally accepted benchmark for CBIR and the lack of a metric is probably one of the main causes for the poor quality of today’s algorithms—without a perfor- mance metric is it impossible to diag- nose the shortcomings of a particular algorithm and identify the critical con- trol points.7 An international effort is underway to create a benchmark for CBIR,8 simi- lar to what was done in the past in the TREC effort for text retrieval. This re- quires an extensive collaboration to an- notate a sufficiently large image cor- pus, which establishes the ground truth Journal of Electronic Imaging / October 2002 / Vol. 11(4) / 421 Downloaded From: https://www.spiedigitallibrary.org/journals/Journal-of-Electronic-Imaging on 05 Apr 2021 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use against which performance can be measured. A tool has recently been de- veloped for this purpose.9 One particularly nasty problem on the internet is that a preponderance of the available images is not normalized towards a standard rendering intent, as is done in conventional stock image collections. In fact, the subtleties of the various references for color encoding in the stages of a distributed workflow are only recently being described and standardized.10 A correct output-referred color en- coding cannot be determined manually in the case of a large image corpus, as it is typically encountered in internet im- aging. Contrary to silver halide photog- raphy, where contemporary films can largely compensate for illumination de- viating from the intended illuminant, this is not the case in digital photogra- phy. This problem has led to the pro- posal of a number of automatic white balancing algorithms to compensate for these discrepancies by estimating the illuminant and applying a color appear- ance transformation. To benchmark these algorithms it is necessary to develop a ground truth for combinations of illuminations and as- sumed illuminants. Tominaga’s paper on a ‘‘Natural image database and its use for scene illumninant estimation’’ describes how such a database is cre- ated and how it is used in practice. Digitalization, compression, and ar- chiving of visual information has be- come popular, inexpensive and straightforward. Yet, the retrieval of this information on the World Wide Web— being highly distributed and minimally indexed—is far from being effective and efficient. A hot research topic is the definition of feasible strategies to mini- mize the semantic gap between the low-level features that can be automati- cally extracted from the visual contents of an image and the human interpreta- tion of such contents. Two different ap- proaches to this problem are described in the last two papers. Lienhart and Hartmann present novel and effective algorithms for clas- sifying images on the web. This type of algorithms will be a key element in the next generation of search engines, which will have to classify the web page media contents automatically. Ex- periments and results are reported and discussed about distinguishing photo- like images from graphical images, ac- tual photos from only photo-like, but ar- tificial images and presentation slides/ scientific posters from comics. The paper ‘‘Multimodal search in collections of images and text’’ by San- tini introduces the intriguing issue of how to infer meaning of an image from both its pictorial content and its context. The author describes a data model and a query algebra for databases of im- ages immersed in the World Wide Web. The author’s model provides a semantic structure that, taking into ac- count the connection with the text and pages containing them, enriches the in- formation that can be recovered from the images themselves. References 1. F.J. Verbeek et al., ‘‘Visualization of com- plex data sets over Internet: 2D and 3D vi- sualization of the 3D digital atlas of zebra fish development,’’ Proc. SPIE 4672, 20–29 ~January 2002!. 2. G. Ciocca et al., ‘‘A multimedia search en- gine with relevance feedback,’’ Proc. SPIE 4672, 243–251 ~January 2002!. 3. G. Beretta, ‘‘WWW1 Structure5 Knowl- edge, ’ ’ Technical Report HPL-96-99, HP Laboratories, Palo Alto, June 1996, http:// www.hpl.hp.com/techreports/96/HPL-96- 99.html 4. J. Tillinghast et al., ‘‘Structure and Naviga- tion for Electronic Publishing,’’ Proc. SPIE 3300, 38 – 45 ~January 1998!. 5. C. Jörgensen et al., ‘‘Testing a vocabulary for image indexing and ground truthing,’’ Proc. SPIE 4672, 207–215 ~January 2002!. 6. D.J. Watts et al., ‘‘Identity and Search in Social Networks,’’ Science 296, 1302–1305 ~May 2002!. 7. N.J. Gunther et al., ‘‘A Benchmark for Im- age Retrieval using Distributed Systems over the Internet: BIRDS-I,’’ Proc. SPIE 4311, 252–267 ~January 2001!. 8. http://www.benchathlon.net/ 9. T. Pfund et al., ‘‘A Dynamic Multimedia Annotation Tool,’’ Proc. SPIE 4672, 216 – 224 ~January 2002!. 10. S. Süsstrunk, ‘‘Color Encodings for Image Databases,’’ Proc. SPIE 4672, 174 –180 ~January 2002!. Giordano Beretta is with the Imaging Systems Labora- tory at Hewlett- Packard. He has been instrumental in bootstrapping the internet imag- ing community: in collaboration with Robert Buckley he has developed a course on ‘‘Color Imaging on the Internet,’’ which they have taught at several IS&T and SPIE conferences; with Raimondo Schettini he has started a series of Internet Imaging conferences at the IS&T/ SPIE Electronic Imaging Symposium; and he has nursed the Benchathlon effort through its first two years. He is a Fellow of the IS&T and the SPIE. Raimondo Schet- tini is an asso- ciate professor at DISCO, University of Milano Bicocca. He has been asso- ciated with Italian National Research Council (CNR) since 1987. In 1994 he moved to the Institute of Multimedia Information Tech- nologies, where he is currently in charge of the Imaging and Vision Lab. He has been team leader in several research projects and published more than 130 refereed pa- pers on image processing, analysis and re- production, and image content-based index- ing and retrieval. He is member of the CIE TC 8/3. He has been General Co-Chairman of the 1st Workshop on Image and Video Content-based Retrieval (1998), and gen- eral co-chair of the EI Internet Imaging Con- ferences (2000-2002). He was general co- chair of the First European Conference on Color in Graphics, Imaging and Vision (CGIV’2002). G U E S T E D I T O R I A L 422 / Journal of Electronic Imaging / October 2002 / Vol. 11(4) Downloaded From: https://www.spiedigitallibrary.org/journals/Journal-of-Electronic-Imaging on 05 Apr 2021 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use