New multiple matched filter: design and experimental realization

A method of making a multiple matched filter which allows the recognition of different characters in successive planes in simple conditions is proposed. The generation of the filter is based on recording on the same plate the Fourier transforms of the different patterns to be recognized, each of which is affected by different spherical phase factors because the patterns have been placed at different distances from the lens. This is proved by means of experiments with a triple filter which allows satisfactory recognition of three characters.

Computation of arbitrarily constrained synthetic discriminant functions

An algorithm for computing correlation filters based on synthetic discriminant functions that can be displayed on current spatial light modulators is presented. The procedure is nondivergent, computationally feasible, and capable of producing multiple solutions, thus overcoming some of the pitfalls of previous methods.

Correlation postprocessing-based method for the detection of de focused images

We present a method to detect patterns in defocused scenes by means of a joint transform correlator. We describe analytically the correlation plane, and we also introduce an original procedure to recognize the target by postprocessing the correlation plane. The performance of the methodology when the defocused images are corrupted by additive noise is also considered.

Reduction of the effect of aberrations in a joint transform correlator

We report the study of the influence of optical aberrations in a joint-transform correlator: The wave aberration of the optical system is computed from data obtained by ray tracing. Three situations are explored: We consider the aberration only in the first diffraction stage (generation of power spectrum), then only in the second (transformation of the power spectrum into correlation), and finally in both stages simultaneously. The results show that the quality of the correlation is determined mostly by the aberrations of the first diffraction stage and that we can optimize the setup by moving the cameras along the optical axis to a suitable position. The good agreement between the predicted data and the experimental results shows that the method explains well the behavior of optical diffraction systems when aberrations are taken into account.

Utilización de pantallas de cristal líquido en un correlador de Vander Lugt

En este artículo se presenta un estudio cuya finalidad es analizar diferentes aspectos de la utilización de pantallas de cristal líquido, extraídas de un videoproyector, en montajes de reconocimiento de formas por correlación. Se analizan las condiciones de funcionamiento de las pantallas y sus posibles modos de configuración. Se estudian dos tipos de filtros de correlación, el filtro adaptado clásico y el de sólo fase, así como la manera de codificarlos en las pantallas. Finalmente, se presentan los resultados de una serie de realizaciones experimentales utilizando un correlador de VanderLugt y diferentes configuraciones de las pantallas. De todo ello se deducen las condiciones óptimas de funcionamiento del sistema.

Sidelobe elimination for generalized synthetic discriminant functions by a two-filter correlation and subsequent postprocessing of the intensity distributions

One of the most important problems in optical pattern recognition by correlation is the appearance of sidelobes in the correlation plane, which causes false alarms. We present a method that eliminate sidelobes of up to a given height if certain conditions are satisfied. The method can be applied to any generalized synthetic discriminant function filter and is capable of rejecting lateral peaks that are even higher than the central correlation. Satisfactory results were obtained in both computer simulations and optical implementation.

Using appearance and context for outdoor scene object classification

We propose a probabilistic object classifier for outdoor scene analysis as a first step in solving the problem of scene context generation. The method begins with a top-down control, which uses the previously learned models (appearance and absolute location) to obtain an initial pixel-level classification. This information provides us the core of objects, which is used to acquire a more accurate object model. Therefore, their growing by specific active regions allows us to obtain an accurate recognition of known regions. Next, a stage of general segmentation provides the segmentation of unknown regions by a bottom-strategy. Finally, the last stage tries to perform a region fusion of known and unknown segmented objects. The result is both a segmentation of the image and a recognition of each segment as a given object class or as an unknown segmented object. Furthermore, experimental results are shown and evaluated to prove the validity of our proposal

Dance movement patterns recognition

"Es tracta d'un projecte dividit en dues parts independents però complementàries, realitzades per autors diferents. Aquest document conté originàriament altre material i/o programari només consultable a la Biblioteca de Ciència i Tecnologia"

Label Space Reduction in All-Optical Label Switchiing Networks

Report for the scientific sojourn at the Department of Information Technology (INTEC) at the Ghent University, Belgium, from january to june 2007. All-Optical Label Swapping (AOLS) forms a key technology towards the implementation of All-Optical Packet Switching nodes (AOPS) for the future optical Internet. The capital expenditures of the deployment of AOLS increases with the size of the label spaces (i.e. the number of used labels), since a special optical device is needed for each recognized label on every node. Label space sizes are affected by the wayin which demands are routed. For instance, while shortest-path routing leads to the usage of fewer labels but high link utilization, minimum interference routing leads to the opposite. This project studies and proposes All-Optical Label Stacking (AOLStack), which is an extension of the AOLS architecture. AOLStack aims at reducing label spaces while easing the compromise with link utilization. In this project, an Integer Lineal Program is proposed with the objective of analyzing the softening of the aforementioned trade-off due to AOLStack. Furthermore, a heuristic aiming at finding good solutions in polynomial-time is proposed as well. Simulation results show that AOLStack either a) reduces the label spaces with a low increase in the link utilization or, similarly, b) uses better the residual bandwidth to decrease the number of labels even more.

Object recognition for autonomous objects: comparison of two approaches

Report for the scientific sojourn at the Swiss Federal Institute of Technology Zurich, Switzerland, between September and December 2007. In order to make robots useful assistants for our everyday life, the ability to learn and recognize objects is of essential importance. However, object recognition in real scenes is one of the most challenging problems in computer vision, as it is necessary to deal with difficulties. Furthermore, in mobile robotics a new challenge is added to the list: computational complexity. In a dynamic world, information about the objects in the scene can become obsolete before it is ready to be used if the detection algorithm is not fast enough. Two recent object recognition techniques have achieved notable results: the constellation approach proposed by Lowe and the bag of words approach proposed by Nistér and Stewénius. The Lowe constellation approach is the one currently being used in the robot localization project of the COGNIRON project. This report is divided in two main sections. The first section is devoted to briefly review the currently used object recognition system, the Lowe approach, and bring to light the drawbacks found for object recognition in the context of indoor mobile robot navigation. Additionally the proposed improvements for the algorithm are described. In the second section the alternative bag of words method is reviewed, as well as several experiments conducted to evaluate its performance with our own object databases. Furthermore, some modifications to the original algorithm to make it suitable for object detection in unsegmented images are proposed.

ICR a partir de traços d'escriptura natural captats amb un bolígraf digital

Aquest projecte consisteix en la realització d’una anàlisi de diferents reconeixedors de caracters manuscrits, concretament de nombres, per a una possible implantació en la digitalització de formularis en la industria. Al llarg del document s’estudien dos reconeixedors diferents, concretament l’incorporat al paquet "Tablet PC and Recognition Pack" de Microsoft i el Heloise Hse, proporcionat per la Universitat de Berkeley a Califòrnia.

Estudi òptic dels minerals

El Grup Consolidat d’Innovació Docent de Mineralogia i òptica cristal·lina de la Universitat de Barcelona ha desenvolupat un CD interactiu que simula el funcionament d’un microscopi petrogràfic, per tal de facilitar a l’alumne un material d’autoaprenentatge, ha de servir per a reforçar els coneixements dels minerals formadors de roques en làmina prima. Aquest material te tres entrades diferents, en català, castellà i anglès. Cada mineral té una fitxa general amb les seves propietats òptiques i una complementaria amb les característiques cristal·logràfiques, camp d’estabilitat, diagrames de fases i característiques morfològiques del mineral a observar, les quals marquen els trets determinatius d’aquell mineral per tal de facilitar el seu reconeixement. Per tal de complementar les dades s’han introduït links directes amb la planes web: “webmineral” i “mindat” on hi ha les corresponents estructures i morfologies “interactives” de cadascun dels minerals que apareixen en el programa. En l’aplicació informàtica hi ha 169 filmacions corresponents a 43 dels principals minerals que formen les roques, una filmació correspon a la imatge només amb el polaritzador, i l'altre a la imatge amb el polaritzador més l'analitzador. Cadascuna d'aquestes imatges es presenta amb un gir de 360º; es pot aturar i després continuar girant, simulant el que veuríem al microscopi. D'aquesta manera es pot determinar el pleocroisme, la presència de macles, el color d'interferència i l'angle d'extinció.. S’ha intentat sempre que hi hagués diferents exemples d’un mateix mineral en diverses paragènesis. També s'incorpora una fitxa que l'usuari pot omplir amb les característiques texturals i òptiques del mineral agrupades segons les observacions que es fan, bé amb el polaritzador, amb el polaritzador i l'analitzador o bé amb les condicions específiques per veure la figura d'interferència i el signe òptic. Aquesta fitxa, un cop plena, es pot imprimir. En tot moment hi ha un menú d’ajuda on l’usuari pot remetre i fer la consulta adient per poder continuar.

Line search multilevel optimization as computational methods for dense optical flow

We evaluate the performance of different optimization techniques developed in the context of optical flowcomputation with different variational models. In particular, based on truncated Newton methods (TN) that have been an effective approach for large-scale unconstrained optimization, we develop the use of efficient multilevel schemes for computing the optical flow. More precisely, we evaluate the performance of a standard unidirectional multilevel algorithm - called multiresolution optimization (MR/OPT), to a bidrectional multilevel algorithm - called full multigrid optimization (FMG/OPT). The FMG/OPT algorithm treats the coarse grid correction as an optimization search direction and eventually scales it using a line search. Experimental results on different image sequences using four models of optical flow computation show that the FMG/OPT algorithm outperforms both the TN and MR/OPT algorithms in terms of the computational work and the quality of the optical flow estimation.

Physically based walking simulation

Hem realitzat l’estudi de moviments humans i hem buscat la forma de poder crear aquests moviments en temps real sobre entorns digitals de forma que la feina que han de dur a terme els artistes i animadors sigui reduïda. Hem fet un estudi de les diferents tècniques d’animació de personatges que podem trobar actualment en l’industria de l’entreteniment així com les principals línies de recerca, estudiant detingudament la tècnica més utilitzada, la captura de moviments. La captura de moviments permet enregistrar els moviments d’una persona mitjançant sensors òptics, sensors magnètics i vídeo càmeres. Aquesta informació és emmagatzemada en arxius que després podran ser reproduïts per un personatge en temps real en una aplicació digital. Tot moviment enregistrat ha d’estar associat a un personatge, aquest és el procés de rigging, un dels punts que hem treballat ha estat la creació d’un sistema d’associació de l’esquelet amb la malla del personatge de forma semi-automàtica, reduint la feina de l’animador per a realitzar aquest procés. En les aplicacions en temps real com la realitat virtual, cada cop més s’està simulant l’entorn en el que viuen els personatges mitjançant les lleis de Newton, de forma que tot canvi en el moviment d’un cos ve donat per l’aplicació d’una força sobre aquest. La captura de moviments no escala bé amb aquests entorns degut a que no és capaç de crear noves animacions realistes a partir de l’enregistrada que depenguin de l’interacció amb l’entorn. L’objectiu final del nostre treball ha estat realitzar la creació d’animacions a partir de forces tal i com ho fem en la realitat en temps real. Per a això hem introduït un model muscular i un sistema de balanç sobre el personatge de forma que aquest pugui respondre a les interaccions amb l’entorn simulat mitjançant les lleis de Newton de manera realista.

Control of optical fields and single photon emitters by advanced nanoantenna structures

The control of optical fields on the nanometre scale is becoming an increasingly important tool in many fields, ranging from channelling light delivery in photovoltaics and light emitting diodes to increasing the sensitivity of chemical sensors to single molecule levels. The ability to design and manipulate light fields with specific frequency and space characteristics is explored in this project. We present an alternative realisation of Extraordinary Optical Transmission (EOT) that requires only a single aperture and a coupled waveguide. We show how this waveguide-resonant EOT improves the transmissivity of single apertures. An important technique in imaging is Near-Field Scanning Optical Microscopy (NSOM); we show how waveguide-resonant EOT and the novel probe design assist in improving the efficiency of NSOM probes by two orders of magnitude, and allow the imaging of single molecules with an optical resolution of as good as 50 nm. We show how optical antennas are fabricated into the apex of sharp tips and can be used in a near-field configuration.