This article is a study of the themes and issues concerning the annotation of digital contents, such as textual documents, images, and multimedia documents in general. These digital contents are automatically managed by different kinds of digital library management systems and more generally by different kinds of information management systems.Even though this topic has already been partially studied by other researchers, the previous research work on annotations has left many open issues. These issues concern the lack of clarity about what an annotation is, what its features are, and how it is used. These issues are mainly due to the fact that models and systems for annotations have only been developed for specific purposes. As a result, there is only a fragmentary picture of the annotation and its management, and this is tied to specific contexts of use and lacks-general validity.The aim of the article is to provide a unified and integrated picture of the annotation, ranging from defining what an annotation is to providing a formal model. The key ideas of the model are: the distinction between the meaning and the sign of the annotation, which represent the semantics and the materialization of an annotation, respectively; the clear formalization of the temporal dimension involved with annotations; and the introduction of a distributed hypertext between digital contents and annotations. Therefore, the proposed formal model captures both syntactic and semantic aspects of the annotations. Furthermore, it is built on previously existing models and may be seen as an extension of them. ACM Reference Format:Agosti M. and Ferro N. 2007. A formal model of annotations of digital content.

Digital libraries (DLs) are new and innovative information systems, under constant development and change, and therefore evaluation is of critical importance to ensure not only their correct evolution but also their acceptance by the user and application communities. The Evaluation activity of the DELOS Network of Excellence has performed a large-scale survey of current DL evaluation activities. This study has resulted in a description of the state of the art in the field, which is presented in this paper. The paper also proposes a new framework for the evaluation of DLs, as well as for recording, describing and analyzing the related research field. The framework includes a methodology for the classification of current evaluation procedures. The objective is to provide a set of flexible and adaptable guidelines for DL evaluation

This paper presents the results of our study regarding the different facets and ways of using annotations in both digital libraries and collaboratories. This study represents an innovative attempt at gathering methodological tools and synergies from both fields in order to effectively define a comprehensive model for annotations. Thus we propose a conceptual model for annotations in order to develop an annotation service that can be plugged into digital libraries and collaboratories. Finally, starting from our model, we introduce a search strategy for exploiting annotations in order to search and retrieve relevant documents for a user query.

This paper discusses how to exploit annotations as a useful context in order to search and retrieve relevant documents for a user query. This paper provides a formal framework which can be useful in facing this problem and shows how this framework can be employed, by using techniques which come from the hypertext information retrieval and data fusion fields

In the last decade, the importance of analyzing information management systems logs has grown, because log data constitute a relevant aspect in evaluating the quality of such systems. A review of 10 years of research on log analysis is presented in this study. About 50 papers and posters from five major conferences and about 30 related journal papers have been selected to trace the history of the state-of-the-art in this field. The study presents an overview of two main themes: Web search engine log analysis and digital library system log analysis. The problem of the analysis of different sources of log data and the distribution of data are investigated

This paper provides a comprehensive study on annotations by defining their contours and complexity. This work adds a new complementary approach to the usual case and user studies, and also investigates history in order to benefit from previous knowledge and our cultural heritage. This study emphasizes an aspect which has never previously been taken into account: the temporal dimension involved in annotations. Moreover, it discusses both the notion of hypertext between documents and annotations and the idea of annotations as context for documents. The study gives the necessary historical and cultural background to derive a set of key features of annotations that must be taken into account when designing systems that have to support the management of digital annotations on digital contents

No abstract

The digitalization of clinical workflows and the increasing performance of deep learning algorithms are paving the way towards new methods for tackling cancer diagnosis. However, the availability of medical specialists to annotate digitized images and free-text diagnostic reports does not scale with the need for large datasets required to train robust computer-aided diagnosis methods that can target the high variability of clinical cases and data produced. This work proposes and evaluates an approach to eliminate the need for manual annotations to train computer-aided diagnosis tools in digital pathology. The approach includes two components, to automatically extract semantically meaningful concepts from diagnostic reports and use them as weak labels to train convolutional neural networks (CNNs) for histopathology diagnosis. The approach is trained (through 10-fold cross-validation) on 3’769 clinical images and reports, provided by two hospitals and tested on over 11’000 images from private and publicly available datasets. The CNN, trained with automatically generated labels, is compared with the same architecture trained with manual labels. Results show that combining text analysis and end-to-end deep neural networks allows building computer-aided diagnosis tools that reach solid performance (micro-accuracy = 0.908 at image-level) based only on existing clinical data without the need for manual annotations.