Papers by Alberto Sanfeliu
Discrete Applied Mathematics, 2005
In this paper a novel method for indexing views of 3D objects is presented. The topological prope... more In this paper a novel method for indexing views of 3D objects is presented. The topological properties of the regions of views of an object or of a set of objects are used to define an index based on region connection calculus and oriented matroid theory. Both are formalisms for qualitative spatial representation and reasoning and are complementary in the sense that, whereas region connection calculus characterize connectivity of couples of connected regions of views, oriented matroids encode relative position of disjoint regions of views and give local and global topological information about their spatial distribution. This indexing technique has been applied to hypothesis generation from a single view to reduce the number of candidates in 3D object recognition processes.
Pattern Recognition, 2006
The aim of this paper is to present a new method to compare histograms. The main advantage is tha... more The aim of this paper is to present a new method to compare histograms. The main advantage is that there is an important time-complexity reduction with respect to the methods presented before. This reduction is statistically and analytically demonstrated in the paper.
Lecture Notes in Computer Science, 2005
In pattern recognition, median computation is an important technique for capturing the important ... more In pattern recognition, median computation is an important technique for capturing the important information of a given set of patterns but it has the main drawback of its exponential complexity. Moreover, the Spectral Graph techniques can be used for the fast computation of the approximate graph matching error, with a considerably reduced execution complexity. In this paper, we merge both
2004 Conference on Computer Vision and Pattern Recognition Workshop, 2004
In this paper we propose a new technique to perform figure-ground segmentation in image sequences... more In this paper we propose a new technique to perform figure-ground segmentation in image sequences of moving objects under varying illumination conditions. Unlike most of the algorithms that adapt color, the assumption of smooth change of the viewing conditions is no longer needed. To cope with this, in this work we introduce a technique that formulates multiple hypotheses about the next state of the color distribution (some of these hypotheses take into account small and gradual changes in the color model and others consider more abrupt and unexpected variations) and the hypothesis that generates the best object segmentation is used to remove noisy edges from the image. This simplifies considerably the final step of fitting a deformable contour to the object boundary, thus allowing a standard snake formulation to successfully track nonrigid contours. Reciprocally, the contour estimation is used to correct the color model. The integration of color and shape is done in a stage denominated 'sample concentration', that has been introduced as a final step to the well-known CONDENSA-TION algorithm.
Lecture Notes in Computer Science, 2003
Non-speech f ound d eature K xtraction > ased on 6 odel z dentification for a obot 8 avigation Ab... more Non-speech f ound d eature K xtraction > ased on 6 odel z dentification for a obot 8 avigation Abstract. Non-speech audio gives important information from the environment that can be used in robot navigation altogether with other sensor information. In this article we propose a new methodology to study non-speech audio signals with pattern recognition techniques in order to help a mobile robot to self-localize in space domain. The feature space will be built with the more relevant coefficients of signal identification after a wavelet transformation preprocessing step given the nonstationary property of this kind of signals.
Lecture Notes in Computer Science, 2004
In order to discriminate different industrial machine sounds contaminated with perturbations (hig... more In order to discriminate different industrial machine sounds contaminated with perturbations (high noise, speech, etc.), a spectral analysis based on a structural pattern recognition technique is proposed. This approach consists of three steps: 1) to de-noise the machine sounds using the Morlet wavelet transform, 2) to calculate the frequency spectrums for these purified signals, and 3) to convert these spectrums into strings, and use an approximated string matching technique, finding a distance measure (the Levenshtein distance) to discriminate the sounds. This method has been tested in artificial signals as well as in real sounds from industrial machines.
International Journal of Pattern Recognition and Artificial Intelligence, 2004
Climbing and Walking Robots: towards New Applications, 2007
International Journal of Pattern Recognition and Artificial Intelligence, 2002
A system that builds and maintains a dynamic map for a mobile robot is presented. A learning rule... more A system that builds and maintains a dynamic map for a mobile robot is presented. A learning rule associated to each observed landmark is used to compute its robustness. The position of the robot during map construction is estimated by combining sensor readings, motion commands, and the current map state by means of an Extended Kalman Filter. The combination of landmark strength validation and Kalman filtering for map updating and robot position estimation allows for robust learning of moderately dynamic indoor environments.
Proceedings. Fourteenth International Conference on Pattern Recognition (Cat. No.98EX170), 1998
Huge and expensive computation resources are usually required to perform graph labelling at high ... more Huge and expensive computation resources are usually required to perform graph labelling at high speed. This fact restricts an extensive use of this methodology in industrial applications such as visual inspection. A new systolic architecture is presented which computes structural distances between cliques of different graphs based on a modified incremental Levenshtein distance algorithm. The distances obtained are used as a support function for graph labelling using probabilistic relaxation techniques. The proposed architecture computes the distances between k input cliques of an input graph and one reference clique of a reference graph. It does not limit the number of cliques nor cliques complexity of the input graph, so any input graph can be labelled. A low cost solution has been implemented based on FPGAs.
Lecture Notes in Computer Science, 1998
In this paper, we present a new fast architecture to compute the distance between cliques in diff... more In this paper, we present a new fast architecture to compute the distance between cliques in different graphs. The distance obtained is used as a support function for graph labelling using probabilistic relaxation techniques. The architecture presented consists on a pipe-lined structure which computes the distance between an input clique and k reference cliques. The number of processing elements needed
World Scientific Series in Computer Science Published 1: Computer-Aided Specification Techniques ... more World Scientific Series in Computer Science Published 1: Computer-Aided Specification Techniques (J Demetrovics, E Knuth & P Rado) 2: Proceedings of the 2nd RIKEN International Symposium on Symbolic and Algebraic Computation by Computers (Eds. N Inada & T Soma) 3: ...
2010 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2010
This article presents a novel approach for optimizing locally the work of cooperative robots and ... more This article presents a novel approach for optimizing locally the work of cooperative robots and obtaining the minimum displacement of humans in a guiding people mission. Unlike other methods, we consider situations where individuals can move freely and can escape from the formation, moreover they must be regrouped by multiple mobile robots working cooperatively. The problem is addressed by introducing a "Discrete Time Motion" model (DTM) and a new cost function that minimizes the work required by robots for leading and regrouping people. The guiding mission is carried out in urban areas containing multiple obstacles and building constraints. Furthermore, an analysis of forces actuating among robots and humans is presented throughout simulations of different situations of robot and human configurations and behaviors.
2010 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2010
This paper describes the validation process of a simulation model that have been used to explore ... more This paper describes the validation process of a simulation model that have been used to explore the new possibilities of interaction when humans are guided by teams of robots that work cooperatively in urban areas. The set of experiments, which have been recorded as video sequences, show a group of people being guided by a team of three people (who play the role of the guide robots). The model used in the simulation process is called Discrete Time Motion model (DTM) described in , where the environment is modeled using a set of potential fields, and people's motion is represented through tension functions. The video sequences were recorded in an urban space of 10.000 m 2 denominated Barcelona Robot Lab, where people move in the urban space following diverse trajectories. The motion (pose and velocity) of people and robots extracted from the video sequences were compared against the predictions of the DTM model. Finally, we checked the proper functioning of the model by studying the position error differences of the recorded and simulated sequences.
ETFA 2001. 8th International Conference on Emerging Technologies and Factory Automation. Proceedings (Cat. No.01TH8597), 2001
This paper describes a method to localize faces in color images based on the fusion of the inform... more This paper describes a method to localize faces in color images based on the fusion of the information gathered from a stereo vision system and the analysis of color images. Our method generates a depth map of the scene and tries to fit a head model taking into account the shape of the model and skin color information. The method is tailored for its use in factory automation applications where the detection and localization of humans is necessary for the completion or interruption of a particular task, such as robot manipulator safety or the interaction of service robots with humans.
Object recognition supported by user interaction for service robots, 2002
A system that performs the tracking of a human head in 3D in real time is presented. The head sha... more A system that performs the tracking of a human head in 3D in real time is presented. The head shape is modeled by an ellipse with a trained color histogram of skin and hair samples. The color histogram is dynamically updated based on incoming image data in order to accommodate for varying illumination conditions. On the other hand, the size of the searched ellipse projected on the image is scaled depending on the depth information gathered from stereo vision. The strength of our method resides on the use of a predictive filter to fuse color and depth information, iteratively refining the location of the head in 3D and the parameters of the head color histogram.
Lecture Notes in Computer Science, 1997
The aim of this work is the characterization of a new structure called Function-DescribedGraphs (... more The aim of this work is the characterization of a new structure called Function-DescribedGraphs (FDG) which can be used to represent objects in computer vision. The FDGs are usefulin synthesizing structural information from a set of objects described through their structure.The FDG nodes and arcs are characterized by the probability distribution of the attributes of theARGs nodes and arcs from
Lecture Notes in Computer Science, 2006
The aim of this paper is to present a new method to compare modulo histograms. In these histogram... more The aim of this paper is to present a new method to compare modulo histograms. In these histograms, the type of elements are cyclic, for instance, the hue in colour images. The main advantage is that there is an important time-complexity reduction respect the methods presented before. The distance between histograms that we present is defined on a structure called
Lecture Notes in Computer Science, 2003
In this paper a novel method for indexing views of 3D objects is presented. The topological prope... more In this paper a novel method for indexing views of 3D objects is presented. The topological properties of the regions of the segmented images of the objects are used to define an index based on oriented matroid theory. Oriented matroids, which are projective invariants, encode incidence relations and relative position of the elements of the image and give local and global topological information about their spatial distribution. This indexing technique is applied to 3D object hypothesis generation from single views to reduce the number of candidates in object recognition processes.
Uploads
Papers by Alberto Sanfeliu