Catálogo de publicaciones - libros

In this paper we present a system, DoLSuD, for the automatic discovery of relevant substructures in a document layout. DoLSuD, Document Layout Substructure Discovery, extracts, analyzes and describes the visual content of structured documents, such as catalogs, in order to discover repeating and distinctive substructures in the document layout and to establish relations between textual and image content. The paper presents the system along with experimental results and the web based service which utilizes the analysis results.

This paper presents an advanced face recognition system based on AdaBoost algorithm in the JPEG compressed domain. First, the dimensionality is reduced by truncating some of the block-based DCT coefficients and the nonuniform illumination variations are alleviated by discarding the DC coefficient of each block. Next, an improved AdaBoost.M2 algorithm which uses Euclidean Distance(ED) to eliminate non-effective weak classifiers is proposed to select most discriminative DCT features from the truncated DCT coefficient vectors. At last, the LDA is used as the final classifier. Experiments on Yale face databases show that the proposed approach is superior to other methods in terms of recognition accuracy, efficiency, and illumination robustness.

To reduce the semantic gap between low-level visual features and the richness of human semantics, this paper proposes new algorithms, by virtue of the combined camera motion descriptors with multi-threshold, to automatically retrieve the semantic concepts, i.e., close-up, and panorama, directly in MPEG compressed domain based on camera motion analysis. Extensive experiments illustrate that the proposed algorithms provide promising retrieval results under real-time application scenario and without human intervention.

We discuss three approaches, 3store, D2R and MySQL, we have explored to support efficient querying of multimedia data sources via mSpace, a rich UI. Our results underline key research challenges facing the development of high performance RDF query layers to support complex real-time UIs.

In order to evaluate image annotation and object categorisation algorithms, ground truth in the form of a set of images correctly annotated with text describing each image is required. Statistics on the WordNet categories of keywords collected from recent automated image annotation and object categorisation publications and evaluation campaigns are presented. These statistics provide a snapshot of keywords used to train and test current image annotation systems as well as information on the usefulness of WordNet for categorising them.

We describe in this poster/short paper on-going work on the extraction and semantic interpretation of text regions in television news programmes. We present some of the data we consider in this work, the actual technologies in use and where they have to be improved. Finally we briefly discuss a possible innovative and valuable approach to the establishment of a cross-media analysis framework.

Content based multimedia retrieval systems have been proposed to allow for automatic and efficient indexing and retrieval of the increasing amount of audiovisual data (image, video and audio clips). The search for specific persons within this data is an important subtopic due to its large range of applications. This article describes an original system for multimodal person search and provides some initial performance results that demonstrate the efficiency of the system.

In this paper, we present a system framework for event detection in pedestrian and tracking applications. The system is built upon a robust computer vision approach to detecting and tracking pedestrians in unconstrained crowded scenes. Upon this framework we propose a pedestrian indexing scheme and suite of tools for detecting events or retrieving data from a given scenario.

Multimedia indexing is about developing techniques allowing people to effectively find media. Content-based methods become necessary when dealing with big databases. Current technology allows exploring the emotional space which is known to carry very interesting semantic information. In this paper we state the need for an integrated method which extracts reliable affective information and attaches this semantic information to the medium itself. We present a list of possible applications and advantages that the emotional information can bring about together with a framework called SAMMI and the preliminary results of this newly initiated research work.