The use of multimedia in all its forms is increasingly present in our lives. Bringing the technology to increasingly larger sectors of the population has prompted a rare availability of multimedia in its many forms (image, video, sound, text). But this availability becomes virtually useless if we do not have tools that provide us with the location of the documents that we are interested. The field of multimedia information retrieval pursues as its basic objective the development of such tools and solutions in this context has become a hot area of research.
This application incorporates the results from previous research in our group (particularly those obtained during the execution of projects Imagina and GIM-TECAL-MRI) on pattern recognition (feature extraction and selection), multidimensional indexing and searching by similarity and learning Automatic (Machine Learning). Our long term goal is to evolve a system Qatris from CBIR (Content-Based Image Retrieval) system to a full-MIR (Multimedia Information Retrieval), able to respond to the storage, indexing and search of any multimedia document.
The development of a MIR system is a project of considerable scope, taking into account the different techniques, methods, strategies and technologies to be put into play to handle manner very different types of documents. Our strategy for achieving long-range goal is to give qualitative leap in the incorporation of successive documents of various kinds. In this project we plan to investigate different strategies for storing, indexing and searching documents in collections of digital video in order to pave the way for incorporating video as a kind of document to be treated by our system Qatris.
The field research project includes vManager since January 1st 2009 until December 31 2011.
The starting point of this project is the existence of the application Qatris iManager, which integrates the techniques and methods (developed or implemented by us) from different disciplines covered by our group, to facilitate the storage, sorting and searching content image collections.
In this project we are investigating strategies for incorporating video as a kind of document to be treated by our system Qatris. While the similarities between Image and video allow us to reproduce documents on video, now incorporated in Qatris techniques, their physical differences (representation, storage, structure) and conceptual (on how the user perceives and uses the information in any means) require study and analysis of new technologies, architectures, models and methods, which will pave the way for the incorporation of video in Qatris.
- Analyze possible strategies for the processing, indexing and searching document collections of digital video with no sound.
- Establish a set of technical and scientific requirements for the effective incorporation papers Qatris video system.
This page contains contact information for all project members. From here you can send messages or visit their personal home pages.
The most important scientific results of this project can be summarized in the following publications:
- T. Alonso, M. Barrena, P.G. Rodríguez, A. Polo, M. Salas, A. Caro, M.L. Durán, F. R. Rodríguez, L. Arévalo, A. Corral, M. Ávila, J. Martínez, F. Ruiz, S. Martín. "GIM at TRECVID 2012. The Light Semantic Indexing Task". TREC Video Retrival Evaluation, TRECVID'12, Gaithersburg, Maryland, USA, Nov. 26-28, 2012. Download .
- Félix R. Rodríguez, Manuel Barrena. "A fast and robust bulk-loading algorithm for indexing very large digital elevation datasets. I. Algorithm". Computers & Geosciences Journal 37 (7), 804–813, doi: 10.1016/j.cageo.2011.01.003. (Impact Factor JCR-2011 = 1,429; JCR-5 year = 1,714. Cathegories: Computer Science, Multi/Interdisciplinary Applications 45/99 2º quartil; Geosciences, Multidisciplinary 78/170 2º quartil).
- Félix R. Rodríguez, Manuel Barrena. "A fast and robust bulk-loading algorithm for indexing very large digital elevation datasets. II. Experimental Results". Computers & Geosciences Journal 37 (7), 814–821, doi: 10.1016/j.cageo.2011.01.004.(Impact Factor JCR-2011 = 1,429; JCR-5 year = 1,714. Cathegories: Computer Science, Multi/Interdisciplinary Applications 45/99 2º quartil; Geosciences, Multidisciplinary 78/170 2º quartil).
- Andrés Caro, Pablo G. Rodríguez, Rubén Morcillo, and Manuel Barrena. "vManager, developing a complete CBVR system". 5th Iberian Conference on Pattern Recognition and Image Analysis (IbPRIA’11), Las Palmas (Spain). Lectures Notes in Computer Science (LNCS), Vol.6669, pp.604-611, Springer-Verlag, 2011. Download
- Rubén Morcillo, Pablo G. Rodríguez, Andrés Caro and Manuel Barrena. "vManager: un sistema CBVR basado en color local". XV Jornadas de Ingeniería del Software y Bases de Datos (JISBD’10), September, 7 - 10, 2010, Valencia, Spain, pp. 163-174. Download
- J. M. Lanza, Miryam Salas, and Manuel Barrena. "vManager: una herramienta para la gestión de videos". XV Jornadas de Ingeniería del Software y Bases de Datos (JISBD’10), September, 7 - 10, 2010, Valencia, Spain, Download