The China-US Million Book Digital Library Project (Million Book Project) is an international cooperation program between China and the US. However, one million digitized books are considered not to be the ultimate goal of the project, but a first step towards universal access to human knowledge. In particular, there are four challenges about the new way to analyze, process, operate, visualize and interact with digital media resource in this library. To tackle these challenges, North China Centre of Million Book Project (in Chinese Academy of Sciences) has initiated several innovative research projects in areas such as multimedia content analysis and retrieval, bilingual services, multimodal information presentation, and knowledge-based organization and services. In this keynote speech, we simply review our work in these areas, and argue that by technological cooperation with these innovation research topics, the project will develop a top-level digital library platform for the million book library.

Advances in computing and communications mean that we can cost-effectively store every book, sound recording, movie, software package, and public Web page ever created, and provide access to these collections via the Internet to students and adults all over the world. By mostly using existing institutions and funding sources, we can build this as well as compensate authors within the current worldwide library budget. We should take advantage of our new technologies and our open societies to make a Universal Library again, and go the next step and make all knowledge easily available to every man, woman and child around the world.

Google’s announcement that it intended to digitize all the books in several major research libraries was met with mixed reactions. John Wilkin at the University of Michigan declared “This is the day the world changes,” while Rory Litwin said in Library Juice that the move would “commercialize the great research libraries with a handshake, suddenly and epochally.” The four directors of the Universal Library and Million Book Project have received many questions about the comparative aspects of our work and Google Print. My purpose is to compare the two, talking about their genesis, the realities of collections and logistics, and the worries that arise from these realities.

A distinguishing edfeature of thata digital library has is that it has possesses Terabyte volumes of multimedia resources. One challenge for researchers in the field of multimedia is to find a testbed for showing whichdemonstratethe potentials of multimedia technologies such as video summarization, semantic annotation, multimedia cross indexing and retrieval, and etcso on. Ddeeper inresearch and wider in applications, of indigital libraries revealed their indispensable role as testbed for , one fact is certain: the digital library is becoming a test bed for multimedia technologies. This paper presents challenging issues of some key techniques usemployed in digital libraries and their , its specific needs for multimedia technologies.

This paper starts with a description of the present status of the Digital Library of India Initiative. As part of this initiative large corpus of scanned text is available in many Indian languages and has stimulated a vast amount of research in Indian language technology briefly described in this paper. Other than the Digital Library of India Initiative which is part of the Million Books to the Web Project initiated by Prof Raj Reddy of Carnegie Mellon University, there are a few more initiatives in India towards taking the heritage of the country to the Web. This paper presents the future directions for the Digital Library of India Initiative both in terms of growing collection and the technical challenges in managing such large collection poses.

Global Memory Net (GMNet) is intended to be an effective gateway to the world cultural, historical, and heritage image collections from selected academic educational and research partners in the world. Much of these unique collections of great value to education and research are not currently accessible due to distance, form, and technical barriers. This project is to find new ways to enable users to access and exploit these significant research collections via global network. As GMNet is ending its first 5-year phase in October 2005, it has contributed substantially to the community building in digital library development by accommodating numerous collaborators and technical staff from various parts of the world to spend 3 to 5 months as a full-member of the GMNet team in Boston. They have come from different parts of China—such as Sichuan, Hainan, Shanghai and Xi’an; Croatia; and Hanoi, Vietnam. In addition to contribute to the overall system development and enhancement of system functionalities, they have brought valuable sample image collections of their own institutions/countries, and actually developed prototype collections as a part of GMNet. This paper describes the exciting and productive experience of the first of this visiting research group in developing the GMNet’s Version 2.0 PHP-based system under Prof. Chen’s overall supervision. It also describes both the system’s technical level structure—user/Web-based application/data, and complex functionalities with multi-collection, multi-lingual, multi-modal searching capabilities; system management capabilities; as well as provisions for user uploads and retrieval for our own projects. This Version 2.0 system is built on the Linux/Apache/PHP/MySQL platform. What is described in this paper is an actual case which has formed a base for further new development by others in the research group. It demonstrates fully the value of the synergistic collaboration among global partners for universal digital library development. More information can be found in http://www.memorynet.org/.

An important area of focus in complex systems development is the capability to adapt to variable runtime environmental resources and to accommodate runtime system failures. The research in this area is broadly termed as “Self Healing” and has recently aroused increasing attention to complex systems. Digital libraries have gained popularity because of the richer features they provide compared to traditional libraries. As more users begin to use digital libraries, addressing downtimes of these valuable resources has become a high priority. Operating and providing access to digital content to anyone, anytime from anywhere in the world, results in a continued rise of administrative overhead for system monitoring and needs continuous human intervention. Given the volume of information and the huge infrastructure of modern libraries, continuous manual system administration is quite costly and not feasible. In this paper we propose a self healing digital library system as the solution to this problem and present the approach of adding self healing capabilities to an existing digital library project, the Digital Library of India (DLI). We also propose a self healing framework that enables successful reuse of our approach to other architecturally similar digital library systems.

The objective of Universal Digital Library (UDL) is to capture all books in digital format. A text to speech (TTS) interface for UDL portal would enable access to the digital content in voice mode, and also provide access to the digital content for illiterate and vision-impaired people. Our work focuses on design and implementation of text to speech interface for UDL portal primarily for Indian languages. This paper is aimed at identifying the issues involved in integrating text to speech system into UDL portal and describes the development process of Hindi, Telugu and Tamil voices under Festvox framework using unit selection techniques. We demonstrate the quality of the Tamil and Telugu voices and lay out the plan for integrating the TTS into the UDL portal.

Protecting and preserving our environmental systems require the ability to understand the spatio-temporal distribution of soils, parent material, topography, and land cover as well as the effects of human activities on ecosystems. Space-time modelling of ecosystems in an environmental digital library is essential for visualizing past, present, and future impacts of changes occurring within such landscapes (e.g., shift in land use practices). In this paper, we describe three novel features, spatio-temporal indexing, visualization, and geostatistical genre, for the environmental digital library, Environmental Visualization and Geographic Enterprise System (ENVISAGE), currently in progress at the University of Florida.

China-America Digital Academic Library Project (CADAL) is a collaborative project between universities and institutes in China and the USA, which aims to provide universal access to large scale digital resources and explore the ways of applying multimedia and virtual reality technologies to digital library. The distinct characteristic of the resources in CADAL is that it not only contains one million digital books of different languages, but also contains Terabyte level multimedia resources (image, video, and so on), which are utilized for education and research purposes. So, in the Portal to CADAL, both the traditional services of browsing and searching of digital books, and the services of quickly retrieving and structurally browsing of multimedia documents should be provided. In addition, the services of visual presentation of retrieved results are required too. In this paper, the underlying novel multimedia retrieval methods as well as visualization techniques, which are used in the CADAL portal, are investigated.

This paper reports part of a study to develop a method for automatic multi-document summarization. The current focus is on dissertation abstracts in the field of sociology. The summarization method uses macro-level and micro-level discourse structure to identify important information that can be extracted from dissertation abstracts, and then uses a variable-based framework to integrate and organize extracted information across dissertation abstracts. This framework focuses more on research concepts and their research relationships found in sociology dissertation abstracts and has a hierarchical structure. A taxonomy is constructed to support the summarization process in two ways: (1) helping to identify important concepts and relations expressed in the text, and (2) providing a structure for linking similar concepts in different abstracts. This paper describes the variable-based framework and the summarization process, and then reports the construction of the taxonomy for supporting the summarization process. An example is provided to show how to use the constructed taxonomy to identify important concepts and integrate the concepts extracted from different abstracts.

As the popularity of digital images is rapidly increasing on the Internet, research on technologies for semantic image classification has become an important research topic. However, the well-known content-based image classification methods do not overcome the so-called semantic gap problem in which low-level visual features cannot represent the high-level semantic content of images. Image classification using visual and textual information often performs poorly since the extracted textual features are often too limited to accurately represent the images. In this paper, we propose a semantic image classification approach using multi-context analysis. For a given image, we model the relevant textual information as its multi-modal context, and regard the related images connected by hyperlinks as its link context. Two kinds of context analysis models, i.e., cross-modal correlation analysis and link-based correlation model, are used to capture the correlation among different modals of features and the topical dependency among images induced by the link structure. We propose a new collective classification model called relational support vector classifier (RSVC) based on the well-known Support Vector Machines (SVMs) and the link-based correlation model. Experiments showed that the proposed approach significantly improved classification accuracy over that of SVM classifiers using visual and/or textual features.

The convergence of next-generation Networks and the emergence of new media systems have made media-rich digital libraries popular in application and research. The discovery of media content objects’ usage patterns, where QPop Increment is the characteristic feature under study, is the basis of intelligent data migration scheduling, the very key issue for these systems to manage effectively the massive storage facilities in their backbones. In this paper, a clustering algorithm is established, on the basis of temporal segmentation of QPop Increment, so as to improve the mining performance. We employed the standard C-Means algorithm as the clustering kernel, and carried out the experimental mining process with segmented QPop Increases obtained in actual applications. The results indicated that the improved algorithm is more advantageous than the basic one in important indices such as the clustering cohesion. The experimental study in this paper is based on a Media Assets Library prototype developed for the use of the advertainment movie production project for Olympics 2008, under the support of both the Humanistic Olympics Study Center in Beijing, and China State Administration of Radio, Film and TV.

In order to support users to search and browse for various resources, digital libraries are composed of discovery systems which provide user interface and information retrieve system. Recent researches in Information Retrieval have investigated different techniques through improving precision and recall to enhance the effectiveness of discovery system in digital libraries. In this paper, we present our work to enhance discovery system effectiveness with a different approach, through resource discovery based on visual knowledge navigation. In our Strand Map Services project under National Science Digital Library, we introduce the visual resource discovery system called conceptual browsing interfaces, to help educators and learners to locate, comprehend and use educational resources in digital libraries. The paper begins with a short introduction of the Strand Map Services. Then we illustrate the service architecture, the design and implementation of its major components. We will focus our discussion of how the visualization system of the Strand Map Services supports the visual knowledge navigation for distributed digital libraries. This includes the knowledge acquisition of the conceptual browsing interfaces, different knowledge representations in the system perspective and user interface perspective, visualization system modules, algorithm and Web services integration to use visual knowledge navigation to enhance resource discovery in digital libraries.

This paper briefly introduces the main ideas of a sustainable development OCR system based on open architecture techniques and then describes the construction of an optical character recognition (OCR) center built on computer clusters, for the purpose of dynamically improving the recognition precision of the digitized texts of a million volumes of books produced by the China-US Million Books Digital Library (CADAL) Project. The practice of this center will provide helpful reference for other digital library projects.

Information seekers are generally on their own to discover and use a research library’s growing array of digital collections, and coordination of these collections’ development and maintenance is often not optimal. The frequent lack of a conscious design for how collections fit together is of equal concern because it means that research libraries are not making the most of the substantial investments they are making in digital initiatives. This paper proposes a framework for a research library’s digital collections that offers integrated discovery and a set of best practices to underpin collection building, federated access, and sustainability. The framework’s purpose is to give information seekers a powerful and easy way to search across existing and future collections and to retrieve integrated sets of results. The paper and its recommendations are based upon research undertaken by the author and a team of librarians and technologists at Cornell University Library. The team conducted structured interviews of forty-five library staff members involved in digital collection building at Cornell, studied an inventory of the library’s more than fifty digital collections, and evaluated seven existing Open Archives Initiative (OAI) and federated search production or prototype systems. The author will discuss her team’s research and the rationale for their recommendations to: present a cohesive view of the library’s digital collections for both browsing and searching at the object level; take a programmatic (rather than project-based) approach to digital collection building; require that all new digital collections conform to library-developed and agreed-upon OAI best practices for data providers; and implement organizational structures to sustain the library’s digital collections over the long term.

The Bibliotheca Alexandrina (BA) has been developing and putting to use a workflow for turning printed books into digital books as its contribution to the building of a Universal Digital Library. This workflow is a process consisting of multiple phases, namely, scanning, image processing, OCR, digital archiving, document encoding, and publishing. Over the past couple of years, the BA has defined procedures and special techniques for the scanning, processing, OCR and publishing, especially of Arabic books. This workflow has been automated, allowing the governance of the different phases and making possible the production of 18000 books so far. The BA has also designed and implemented a framework for the encoding of digital books that allows publishing as well as a software system for managing the creation, maintenance, and publishing of the overall digital repository.

In the last two decades of the 20th century, there has been an increasing interest in and emphasis on the study of the Hong Kong literature in both the academic and general public in Hong Kong. Recognizing the emergent need of the resources on Hong Kong literature, the University Library System of the Chinese University of Hong Kong set up the Hong Kong Literature Database (the “Database”), which was the first Chinese literature database in the Internet in 2000. The paper will examine how the database is constructed using XML technology and metadata schema. The database also employs Unicode UTF-8 as the internal code. A mapping table for traditional and simplified Chinese characters was created based on Unihan and is used behind the scene so that a user can either input traditional or simplified Chinese characters and retrieval will give both traditional and simplified Chinese characters. Currently 65% of journals use OCR technology so that full-text searching is possible. The Chinese OCR technology will be examined in greater detail. Special features of the Database such as, page-by-page browse mode, position-highlight for full-page newspaper, linking Table-Of-Contents and book jackets from the Library catalogue, etc. are described. The paper will also bring out the problem of massive downloading and compare the state-of-the-art technology and their shortcomings. This paper shows how the Hong Kong Literature Database facilitates future collaboration and data exchange by using open standard, shareable structure and the latest technology.

Many different languages are spoken in India, each language being the mother tongue of tens of millions of people. While the languages and scripts are distinct from each other, the grammar and the alphabet are similar to a large extent. One common feature is that all the Indian languages are phonetic in nature. In this paper we describe the development of a transliteration scheme Om which exploits this phonetic nature of the alphabet. Om uses ASCII characters to represent Indian language alphabets, and thus can be read directly in English, by a large number of users who cannot read script in other Indian languages than their mother tongue. It is also useful in computer applications where local language tools such as email and chat are not yet available. Another significant contribution presented in this paper is the development of a text editor for Indian languages that integrates the Om input for many Indian languages into a word processor such as Microsoft WinWord^®. The text editor is also developed on Java^® platform that can run on Unix machines as well. We propose this transliteration scheme as a possible standard for Indian language transliteration and keyboard entry.

Transliteration editors are essential for keying-in Indian language scripts into the computer using QWERTY keyboard. Applications of transliteration editors in the context of Universal Digital Library (UDL) include entry of meta-data and dictionaries for Indian languages. In this paper we propose a simple approach for building transliteration editors for Indian languages using Unicode and by taking advantage of its rendering engine. We demonstrate the usefulness of the Unicode based approach to build transliteration editors for Indian languages, and report its advantages needing little maintenance and few entries in the mapping table, and ease of adding new features such as adding letters, to the transliteration scheme. We demonstrate the transliteration editor for 9 Indian languages and also explain how this approach can be adapted for Arabic scripts.

The principles and methods of TQM (total quality management) were applied to digital library management. TQM puts emphasis on process control and continual improvement. The TQM theories and measures explained clearly in this paper were the scientific tools used in the experiment in the Wenzhou University Digital Library. This paper deals with a very broad range of subjects related to a digital library: system model of TQM for digital library, organization structure, and tasks of Wenzhou University Digital Library, task flow chart of electronic reading room, cause and effect of electronic reading room quality management, systematic model of enhancing electronic information acquisition rate, quality comparison of electronic reading room service, etc.

Digitization projects should focus on quantity rather than quality. Increasing quantities of information produce qualitatively more valuable services. Online writing and searching are now common, and it is only online reading that is still limiting our use of online books. New interfaces might increase our willingness to read online, which should be encouraged rather than fought, since it represents an increase both the amount of information available and the participation of more people in the writing and exchange of information.

Copyright and its international complications have presented a significant barrier to the Universal Digital Library (UDL)’s mission to digitize all the published works of mankind and make them available throughout the world. We discuss the effect of existing copyright treaties and various proposals, such as compulsory licensing and the public lending right that would allow access to copyrighted works without requiring permission of their owners. We argue that these schemes are ineffective for purposes of the UDL. Instead, making use of the international consensus that copyright does not protect facts, information or processes, we propose to scan works digitally to extract their intellectual content, and then generate by machine synthetic works that capture this content, and then translate the generated works automatically into multiple languages and distribute them free of copyright restriction.

In this digital era, we have witness the exciting convergence of content, technology, and global collaboration in the development of digital libraries. The mode of universal access for information seeking and knowledge acquisition differs greatly from the traditional ways. From the information resources point of views, the old model of “owning” a collection has given way to “sharing,” and the new emphases have shifted from possessing large “physical libraries” to “virtual libraries” digitally distributed all over the world. “Universal access” has taken on a very different meaning when one has the ability to share invaluable resources through the use of cutting edge technologies. The author has experienced much of these transformations through her own R&D activities—from the creation of interactive videodisc and multimedia CD on the First Emperor of China’s terracotta warriors and horses in the 1980s and 1990s to leading a current international digital library project, Global Memory Net (GMNet), supported by the US National Science Foundation. In presenting her vision for linking world digital resources together for universal access, she will share with the audience the latest development of Global Memory Net.

With the development of digital library technology, library books made of paper can be digital released and read, and Endangered Cultural Heritages can be preserved. Traditional library’s contents and functions can be greatly enhanced by digital technologies. For these new library objects, the primary key problem is precisely reconstructing their 3D models. When constructing complete 3D models, multiple color texture maps are often necessary. A commonly encountered problem uncounted during fusing of textures from multiple color images is color distortion. Each texture of a single 3D model may be obtained under possibly different lighting conditions and color response of the camera. To remove any visible seam and improve color consistency between the textures while avoiding color distortion, we propose a new efficient algorithm to relight all the texture images globally, spread residual light difference, and recolor each image by homogeneous transformation. A relative illumination model was adopted to obtain the relighting function. We choose lαβ color space with minimal correlation between channels for many natural scenes, for calculating the relighting result. Looking into two overlapped images A and B, we can pairwise relight B into A’s luminosity condition in two steps. We first scale B’s l channel by the l_A/l_B ratio of the overlapped region. We can assume A and B are in a same color plane now. Then a homogeneous transformation is applied to B’s α and β channels which moves B into A’s hue and saturation condition. For multiple overlapped color textures, a patch based weighted global relighting method was proposed to minimize the total color difference. The pairwise relighting method was used between each two overlapped images, and the difference in every overlapped region after relighting was weighted and summed up to construct an energy value. We used Nelder-Mead method to find a minimal energy value and the relighting parameters for every image. After global relighting, textures become almost coherent. We simply blended the overlapped region along the texture border to remove small visual seams and get a final result. We illustrate our method by calibrating textures of a painted sculpture acquired with laser scanner. Experimental results were realistic and reliable and showed how this method can fuse multiple textures without color distortion.

Optical Character Recognition (OCR) refers to the process of converting printed Tamil text documents into software translated Unicode Tamil Text. The printed documents available in the form of books, papers, magazines, etc. are scanned using standard scanners which produce an image of the scanned document. As part of the preprocessing phase the image file is checked for skewing. If the image is skewed, it is corrected by a simple rotation technique in the appropriate direction. Then the image is passed through a noise elimination phase and is binarized. The preprocessed image is segmented using an algorithm which decomposes the scanned text into paragraphs using special space detection technique and then the paragraphs into lines using vertical histograms, and lines into words using horizontal histograms, and words into character image glyphs using horizontal histograms. Each image glyph is comprised of 32×32 pixels. Thus a database of character image glyphs is created out of the segmentation phase. Then all the image glyphs are considered for recognition using Unicode mapping. Each image glyph is passed through various routines which extract the features of the glyph. The various features that are considered for classification are the character height, character width, the number of horizontal lines (long and short), the number of vertical lines (long and short), the horizontally oriented curves, the vertically oriented curves, the number of circles, number of slope lines, image centroid and special dots. The glyphs are now set ready for classification based on these features. The extracted features are passed to a Support Vector Machine (SVM) where the characters are classified by Supervised Learning Algorithm. These classes are mapped onto Unicode for recognition. Then the text is reconstructed using Unicode fonts.