We would like to welcome Raúl Crespo to the FRAV research group. Raúl graduated at the University of León (Spain) as a computer engineer and has joined the group to work for CABINTEC Project. Therefore he fills the gap left by Alicia Ballesteros and Víctor López, who left in the end of 2008.
During the meeting of the Universidad Rey Juan Carlos Government Council held on December 19, 2008 (point 8 of the agenda) the following decision was passed unanimously:
"Approval of the following nominations as Honorary Collaborators, as proposed by the Deparment of Computer Architecture and Technology, Computer Science and Artificial Intelligence:
FRAV would like to thank the Universidad Rey Juan Carlos for this courtesy with these three researchers, with whom we share common interests and a close relationship. The Fuentes Covarrubias Brothers belong to the Universidad de Colima (located at the city and state of the same name, Mexico), while Professor Gerardo Reyes works at CENIDET (Cuernavaca, Morelos, Mexico).
Dr Enrique Cabello, FRAV principal investigator, has become a tenured, full time Professor ("Profesor Titular de Universidad"), strengthening definitely his position as a civil servant at Universidad Rey Juan Carlos. The nomination, made in the Official Journal of Spain on 09/24/2008, is effective since the beginning of this month.
Dr Cabello has fulfilled several roles as a professor since 1990, as Associate Professor, Temporary Professor of University School and Professor of University School, first in the University of Salamanca and later at the Universidad Rey Juan Carlos. He received his PhD with cum laude at the Universidad Politécnica de Madrid in 2004 with his thesis about neural networks based face recognition, for which he was also awarded in 2006 as the best thesis of the year. He has led a great amount of research projects and is responsible of a long list of publications.
During September, Dr Gerardo Reyes Salgado will be staying at Universidad Rey Juan Carlos (URJC) with FRAV group. This stay is supported by the Spanish Ministry of Education, Social Policy and Sport, as well as by the academic group of Artificial Intelligence of CENIDET . This is Dr Reyes' the second stay after that of April 2007, and serves as a follow-up of the collaboration agreement signed by URJC and CENIDET in 2007. Dr Reyes has written a proposal with FRAV for the creation of a research network involving Mexican and European institutions and firms for a road safety-driven project relying on artificial intelligence, computer vision and cognitive sciences.
Parallelly to Dr Reyes' stay, a student of CENIDET IA group, Óscar Gerardo Sánchez Siordia (CENIDET 2007 promotion) has begun a four-month stay at URJC as a part of his Master thesis under the supervision of Dr Reyes and Dr Enrique Cabello Pardos (FRAV principal investigator). Mr. Sánchez Siordia's stay began in September and will end late in December this year and will serve to obtain information and process data from a truck cabin simulator based in the north of Spain, and accessed by FRAV. Moreover, during this stay he is expected to write several papers to be submitted to high-level conferences. This stay has receibed financial support of the mixed grants program from CONACYT and the Spanish Ministry of Science and Innovation
FRAV research group from Rey Juan Carlos University has developed a computer-assisted video-surveillance system devoted to the identification of suspicious behaviour at airports. Specifically, the project led by Professor Enrique Cabello has consisted in the development of a computer vision system able to warn the airport security staff in case of finding an abandoned object. The tests have been carried out with real images taken at Terminal 4 Madrid-Barajas International Airport, with the collaboration of the security staff, the Guardia Civil, AENA (institution that manages the Spanish airports) and the Universitary Institute for the Research on Interior Security (UNED).
The project, which is an extension of a previous one related to face recognition at aiports, is a model prototype of a computer-assisted surveillance system. The software receives a video, analyses it and then provides real-time visual results to the operator in charge, as well as a report of all suspicious behaviour detected. The operator must check the truthfulness of this alarm in order to proceed to apply the adequate security protocol.
The system works in five stages. In the first one (background subtraction) an special image showing exclusively the moving objects is obtained. The second one (image thresholding) transforms this image into a noise-subtracted matrix. In the third one (labelling), the segmented independent components are numbered and labelled. The fourth one (tracking of moving objects) takes into account the moving objects present in two images, and displays their speed, acceleration, and so on. Finally the last stage detects the abandoned elements.
The system understands that an object has been abandoned when it does not belong to the place, this is, it is not contained in the background image, and when someone has left it there. In other words, when it has been detected in a previous component. In that case, the abandoned components are then clustered together in a special structure. An abandoned-labelled component does not lose its "abandoned" label easily. Therefore, for instance, the program is able to undestand that a suitcase is still abandoned when a group of people walks in front of it. It only changes its state if no component with similar characteristics is detected in the nearby for a long time.
The obtained results are, according to the researchers, very satisfactory, as the system is, in general, very stable and robust under different situations, provided the recording is of high quality. However several factors have been identified to have an influence in the processing: changes of illumination in the scenes, changes of position of the camera, possible changes of a scene, such as the presence of an advertising stand, zoom or the crowdedness of the area.