Graphic summary
  • Show / hide key
  • Information


Scientific and technological production
  •  

1 to 50 of 364 results
  • Monocular Depth Estimation in Images and Sequences Using Occlusion Cues

     Palou Visa, Guillem
    Defense's date: 2014-02-21
    Department of Signal Theory and Communications, Universitat Politècnica de Catalunya
    Theses

    Read the abstract Read the abstract  Share Reference managers Reference managers Open in new window

    Quan els humans observen una escena, son capaços de distingir perfectament les parts que la composen i organitzar-lesespacialment per tal de poder-se orientar. Els mecanismes que governen la percepció visual han estat estudiats des delsprincipis de la neurociència, però encara no es coneixen tots els processos biològic que hi prenen part. En situacions normals,els humans poden fer servir tres eines per estimar l'estructura de l'escena. La primera és l'anomenada divergència. Aprofita l'úsde dos punts de vista (els dos ulls) i és capaç de determinar molt acuradament la posició dels objectes ,que a una distància defins a cent metres, romanen enfront de l'observador. A mesura que augmenta la distància o els objectes no es troben en el campde visió dels dos ulls, altres mecanismes s'han d'utilitzar. Tant l'experiència anterior com certs indicis visuals s'utilitzen enaquests casos i, encara que la seva precisió és menor, els humans aconsegueixen quasibé sempre interpretar bé el seu entorn.Els indicis visuals que aporten informació de profunditat més coneguts i utilitzats són, per exemple, la perspectiva, les oclusionso el tamany de certs objectes. L'experiència anterior permet resoldre situacions vistes anteriorment com ara saber quins regionscorresponen al terra, al cel o a objectes.Durant els últim anys, quan la tecnologia ho ha permès, s'han intentat dissenyar sistemes que interpretessin automàticamentdiferents tipus d'escena. En aquesta tesi s'aborda el tema de l'estimació de la profunditat utilitzant només un punt de vista iindicis visuals d'oclusió. L'objectiu del treball es la detecció d'aquests indicis i combinar-los amb un sistema de segmentació pertal de generar automàticament els diferents plans de profunditat presents a una escena. La tesi explora tant situacionsestàtiques (imatges fixes) com situacions dinàmiques, com ara trames dins de seqüències de vídeo o seqüències completes. Enel cas de seqüències completes, també es proposa un sistema automàtic per reconstruir l'estructura de l'escena només ambinformació de moviment. Els resultats del treball son prometedors i competitius amb la literatura del moment, però mostrenencara que la visió per computador té molt marge de millora respecte la presició dels humans.

  • Image Segmentation Evaluation and Its Application to Object Detection

     Pont Tuset, Jordi
    Defense's date: 2014-02-19
    Department of Signal Theory and Communications, Universitat Politècnica de Catalunya
    Theses

    Read the abstract Read the abstract  Share Reference managers Reference managers Open in new window

    Les primeres parts d'aquesta Tesi se centren en l'estudi de l'avaluació supervisada d'algorismes de segmentació d'imatges. Supervisada en el sentit que els resultats de la segmentació es comparen amb una anotació feta per un humà, coneguda com a ground truth, per mitjà de diferents mesures de similitud. L'avaluació depèn, per tant, de tres punts principals.En primer lloc, les tècniques de segmentació d'imatge que s'avaluen. Es revisa l'estat de l'art en la segmentació de la imatge, fent una diferència explícita entre aquestes tècniques que proporcionen una sortida plana, és a dir, una única agrupació del conjunt de píxels en regions, i les que produeixen una segmentació jeràrquica, és a dir, una estructura en forma d'arbre que representa les regions a diferents escales, des dels detalls fins la imatge completa.En segon lloc, les bases de dades de ground truth són de summa importància en l'avaluació. Es poden dividir en les anotades només a nivell d'objecte, és a dir, amb conjunts de píxels marcats que es refereixen a objectes que no cobreixen tota la imatge; o les que tenen particions completes anotades, que proporcionen una agrupació completa de tots els píxels d'una imatge. Depenent del tipus de base de dades es diu que l'anàlisi es fa des d'una perspectiva objecte o des d'una perspectiva de partició.Finalment, les mesures de similitud utilitzades per comparar els resultats generats amb el ground truth és el que ens proporcionarà una eina quantitativa per avaluar si els resultats són bons, i de quina manera es poden millorar. Les principals aportacions de les primeres parts de la tesi són en el camp de les mesures de similitud.En primer lloc, des d'un punt de vista d'objecte, es revisen les mesures bàsiques que s'utilitzen per comparar dues representacions d'objectes i es demostra que algunes d'elles són equivalents. Per tal d'avaluar les particions completes i jerarquies contra un objecte, cal seleccionar quines de les seves regions formen l'objecte a avaluar. Revisem i millorem aquestes tècniques per mitjà d'un model matemàtic del problema. Aquesta anàlisi ens permet mostrar que les jerarquies poden representar objectes molt millor amb molt menys nombre de regions que les particions planes.Des de la perspectiva de partició, la literatura sobre les mesures d'avaluació és extensa i complexa. La nostra primera contribució és revisar, estructurar i deduplicar les mesures disponibles. Presentem una nova mesura que es demostra que millora les anteriors en termes d'un conjunt de meta-mesures qualitatives i quantitatives. També estenem les mesures sobre les particions planes per avaluar segmentacions jeràrquiques.L'última part d'aquesta Tesi progressa des de l'avaluació de la segmentació d'imatges cap a la seva aplicació a la detecció d'objectes. En particular, ens basem en algunes de les conclusions extretes en la primera part per generar candidats d'objectes segmentats. Donat un conjunt de jerarquies, construïm els parells i triplets de regions, aprenem a combinar el conjunt de cada jerarquia i a prioritzar-los segons descriptors de baix i mig nivell. Duem a terme una extensa validació experimental que mostra que el nostre mètode supera l'estat de l'art en moltes mètriques analitzades.

  • Monocular depth ordering using T-Junctions and convexity occlusion cues

     Palou Visa, Guillem; Salembier Clairon, Philippe Jean
    IEEE transactions on image processing
    Date of publication: 2013
    Journal article

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • Bilateral distance based filtering for polarimetric SAR data

     Alonso Gonzalez, Alberto; López Martínez, Carlos; Salembier Clairon, Philippe Jean; Deng, Xinping
    Remote Sensing
    Date of publication: 2013-10-30
    Journal article

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    This paper introduces a non-linear Polarimetric SAR data filtering approach able to preserve the edges and small details of the data. It is based on exploiting the data locality in both, the spatial and the polarimetric domains, in order to avoid mixing heterogeneous samples of the data. A weighted average is performed over a given window favoring pixel values that are close on both domains. The filtering technique is based on a modified bilateral filtering, which is defined in terms of spatial and polarimetric distances. These distances encapsulate all the knowledge in both domains for an adaptation to the data structure. Finally, the proposed technique is employed to process a real RADARSAT-2 dataset.

  • PolSAR time series processing and analysis based on Binary Partition Trees

     Alonso Gonzalez, Alberto; López Martínez, Carlos; Salembier Clairon, Philippe Jean
    International Workshop on Science and Applications of SAR Polarimetry and Polarimetric Interferometry
    Presentation's date: 2013-02-12
    Presentation of work at congresses

     Share Reference managers Reference managers Open in new window

  • Hierarchical video representation with trajectory binary partition tree

     Palou Visa, Guillem; Salembier Clairon, Philippe Jean
    IEEE Conference on Computer Vision and Pattern Recognition
    Presentation's date: 2013-06-20
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    As early stage of video processing, we introduce an iter- ative trajectory merging algorithm that produces a region- based and hierarchical representation of the video se- quence, called the Trajectory Binary Partition Tree (BPT). From this representation, many analysis and graph cut tech- niques can be used to extract partitions or objects that are useful in the context of specific applications. In order to define trajectories and to create a precise merging algorithm, color and motion cues have to be used. Both types of informations are very useful to characterize objects but present strong differences of behavior in the spa- tial and the temporal dimensions. On the one hand, scenes and objects are rich in their spatial color distributions, but these distributions are rather stable over time. Object mo- tion, on the other hand, presents simple structures and low spatial variability but may change from frame to frame. The proposed algorithm takes into account this key difference and relies on different models and associated metrics to deal with color and motion information. We show that the proposed algorithm outperforms existing hierarchical video segmentation algorithms and provides more stable and pre- cise regions

  • Ensemble learning and hierarchical data representation for microarray classification

     Bosio, Mattia; Bellot, Pau; Salembier Clairon, Philippe Jean; Oliveras Verges, Albert
    IEEE International Conference on Bioinformatics and Bioengineering
    Presentation's date: 2013-11-11
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    The microarray data classification is an open and active research field. The development of more accurate algorithms is of great interest and many of the developed techniques can be straightforwardly applied in analyzing different kinds of omics data. In this work, an ensemble learning algorithm is applied within a classification framework that already got good predictive results. Ensemble techniques take individual experts, (i.e. classifiers), to combine them to improve the individual expert results with a voting scheme. In this case, a thinning algorithm is proposed which starts by using all the available experts and removes them one by one focusing on improving the ensemble vote. Two versions of a state of the art ensemble thinning algorithm have been tested and three key elements have been introduced to work with microarray data: the ensemble cohort definition, the nonexpert notion, which defines a set of excluded expert from the thinning process, and a rule to break ties in the thinning process. Experiments have been done on seven public datasets from the Microarray Quality Control study, MAQC. The proposed key elements have shown to be useful for the prediction performance and the studied ensemble technique shown to improve the state of the art results by producing classifiers with better predictions.

  • Object recognition in urban hyperspectral images using binary partition tree representation

     Valero, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    IEEE International Geoscience and Remote Sensing Symposium
    Presentation's date: 2013-07-18
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    In this work, an image representation based on Binary Partition Tree is proposed for object detection in hyperspectral images. The BPT representation defines a search space for constructing a robust object identification scheme. Spatial and spectral information are integrated in order to analyze hyperspectral images with a region-based perspective. Experimental results demonstrate the good performances of this BPT-based approach.

  • Hierarchical clustering combining numerical and biological similarities for gene expression data classification

     Bosio, Mattia; Salembier Clairon, Philippe Jean; Bellot Pujalte, Pau; Oliveras Verges, Albert
    IEEE Engineering in Medicine and Biology Society
    Presentation's date: 2013-07
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    High throughput data analysis is a challenging problem due to the vast amount of available data. A major concern is to develop algorithms that provide accurate numerical predictions and biologically relevant results. A wide variety of tools exist in the literature using biological knowledge to evaluate analysis results. Only recently, some works have included biological knowledge inside the analysis process improving the prediction results.

  • From local occlusion cues to global monocular depth estimation

     Palou Visa, Guillem; Salembier Clairon, Philippe Jean
    IEEE International Conference on Acoustics, Speech, and Signal Processing
    Presentation's date: 2012-03-14
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    In this paper, we propose a system to obtain a depth ordered seg- mentation of a single image based on low level cues. The algorithm first constructs a hierarchical, region-based image representation of the image using a Binary Partition Tree (BPT). During the building process, T-junction depth cues are detected, along with high convex boundaries. When the BPT is built, a suitable segmentation is found and a global depth ordering is found using a probabilistic framework. Results are compared with state of the art depth ordering and figure/ground labeling systems. The advantage of the proposed ap- proach compared to systems based on a training procedure is the lack of assumptions about the scene content. Moreover, it is shown that the system outperforms previously low-level cue based systems, while offering similar results to a priori trained figure/ground label- ing algorithms

  • Depth ordering on image sequences using motion occlusions

     Palou Visa, Guillem; Salembier Clairon, Philippe Jean
    IEEE International Conference on Image Processing
    Presentation's date: 2012-10-01
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    This paper proposes a system to obtain the depth order of frames in image sequences using motion occlusion cues. The system first computes the forward and backward flows with the previous and next frames and estimates the occluded points. To obtain a region representation of the image, a Binary Partition Tree (BPT) is created for each frame. To estimate occlusion relations in the image, projective flow models are fitted to all regions in the image. The depth order solution is obtained by minimizing over the tree structure a cost function based on occlusion relations and the number of regions. Results show that optical flow algorithms can be used directly to estimate occlusion points. Promising results are obtained combining motion occlusions and region information by means of a BPT. Evaluation is performed comparing current state-of-the-art algorithms on figure/ground assignments, showing that the performance of the proposed system is comparable to current algorithms.

  • Multiclass cancer-microarray classification algorithm with Pair-Against-All redundancy

     Bosio, Mattia; Bellot Pujalte, Pau; Salembier Clairon, Philippe Jean; Oliveras Verges, Albert
    IEEE International Workshop on Genomic Signal Processing and Statistic
    Presentation's date: 2012-12-03
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    Multiclass cancer classification is still a challenging task in the field of machine learning. A novel multiclass approach is proposed in this work as a combination of multiple binary classifiers. It is an example of Error Correcting Output Codes algorithms, applying data transmission coding techniques to improve the classification as a combination of binary classifiers. The proposed method combines the One Against All, OAA, approach with a set of classifiers separating each class-pair from the rest, called Pair Against All, PAA. The OAA+PAA approach has been tested on seven publicly available datasets. It has been compared with the common OAA approach and with state of the art alternatives. The obtained results showed how the OAA+PAA algorithm consistently improves the OAA results, unlike other ECOC algorithms presented in the literature.

  • Microarray classification with hierarchical data representation and novel feature selection criteria

     Bosio, Mattia; Bellot Pujalte, Pau; Salembier Clairon, Philippe Jean; Oliveras Verges, Albert
    IEEE International Conference on BioInformatics and BioEngineering
    Presentation's date: 2012-11-12
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    Microarray data classification is a challenging problem due to the high number of variables compared to the small number of available samples. An effective methodology to output a precise and reliable classifier is proposed in this work as an improvement of the algorithm in [1]. It considers the sample scarcity problem and the lack of data structure typical of microarrays. Both problem are assessed by a two-step approach applying hierarchical clustering to create new features called metagenes and introducing a novel feature ranking criterion, inside the wrapper feature selection task. The classification ability has been evaluated on 4 publicly available datasets from Micro Array Quality Control study phase II (MAQC) classified by 7 different endpoints. The global results have showed how the proposed approach obtains better prediction accuracy than a wide variety of state of the art alternatives.

    Microarray data classification is a challenging prob- lem due to the high number of variables compared to the small number of available samples. An effective methodology to output a precise and reliable classifier is proposed in this work as an improvement of the algorithm in [1]. It considers the sample scarcity problem and the lack of data structure typical of microarrays. Both problem are assessed by a two-step approach applying hierarchical clustering to create new features called metagenes and introducing a novel feature ranking criterion, inside the wrapper feature selection task. The classification ability has been evaluated on 4 publicly available datasets from Micro Array Quality Control study phase II (MAQC) classified by 7 different endpoints. The global results have showed how the proposed approach obtains better prediction accuracy than a wide variety of state of the art alternatives

  • Depth estimation of frames in image sequences using motion occlusions

     Palou Visa, Guillem; Salembier Clairon, Philippe Jean
    European Conference on Computer Vision
    Presentation's date: 2012
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    This paper proposes a system to depth order regions of a frame belonging to a monocular image sequence. For a given frame, re- gions are ordered according to their relative depth using the previous and following frames. The algorithm estimates occluded and disoccluded pixels belonging to the central frame. Afterwards, a Binary Partition Tree (BPT) is constructed to obtain a hierarchical, region based repre- sentation of the image. The nal depth partition is obtained by means of energy minimization on the BPT. To achieve a global depth ordering from local occlusion cues, a depth order graph is constructed and used to eliminate contradictory local cues. Results of the system are evaluated and compared with state of the art gure/ground labeling systems on several datasets, showing promising results.

  • Variable local weight filtering for PolSAR data speckle noise reduction

     Alonso Gonzalez, Alberto; López Martínez, Carlos; Salembier Clairon, Philippe Jean
    IEEE International Geoscience and Remote Sensing Symposium
    Presentation's date: 2012-07
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    This paper presents a Polarimetric SAR data speckle filtering technique, based on a combined filtering in the spatial and polarimetric domains. It is based on a bilateral filtering employing distance measures over these domains. These measures concentrate all the information related to the domain structure that is needed for an adaptation to the scene morphology. A weighted average is performed over a given window favoring closer and similar pixels. As a consequence, an adaptive filtering is achieved, attaining higher filtering over homogeneous areas whereas point scatters remain almost unchanged. Results will be shown over a real RADARSAT-2 data.

  • Temporal PolSAR image series exploitation with binary partition trees

     Alonso Gonzalez, Alberto; López Martínez, Carlos; Salembier Clairon, Philippe Jean
    IEEE International Geoscience and Remote Sensing Symposium
    Presentation's date: 2012-07
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    In this paper, the processing of temporal PolSAR image series is addressed through a region-based and multi-scale data representation, the Binary Partition Tree (BPT). This structure contains useful information related to the data structure at different detail levels that may be employed for different applications. The construction of this structure ans its exploitation is addressed in this work in the context of the speckle filtering and data segmentation applications. A new region model and processing strategy are defined to tackle with the temporal dimension of the data. Finally, to illustrate the capabilities of the proposed technique, results are shown with a real RADARSAT-2 dataset.

  • Processing multidimensional SAR and hyperspectral images with binary partition tree

     Alonso Gonzalez, Alberto; Valero, Silvia; Chanussot, Jocelyn; López Martínez, Carlos; Salembier Clairon, Philippe Jean
    Proceedings of the IEEE
    Date of publication: 2012-08-13
    Journal article

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    The current increase of spatial as well as spectral resolutions of modern remote sensing sensors represents a real opportunity for many prac tical applications but also generates important challenges in terms of image processing. In particular, the spatial correlation between pixels and/or the spectral correlation between spectral bands of a given pixel cannot be ignored. The traditional pixel-based representation of images does not facilitate the handling of these correlations. In this paper, we discuss the inter est of a particular hierarchical region-based representation of images based on binary partition tree (BPT). This representation approach is very flexible as it can be applied to any type of image. Here both optical and radar images will be discussed. Moreover, once the image representation is computed, it can be used for many different applications. Filtering, segmentation, and classifica- tion will be detailed in this paper. In all cases, the interest of the BPT representation over the classical pixel-based representa- tion will be highlighted

  • Hyperspectral image representation and processing with binary partition trees

     Valero Valbuena, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    IEEE transactions on image processing
    Date of publication: 2012-12-04
    Journal article

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    The optimal exploitation of the information provided by hyperspectral images requires the development of advanced image-processing tools. This paper proposes the construction and the processing of a new region-based hierarchical hyperspectral image representation relying on the binary partition tree (BPT). This hierarchical region-based representation can be interpreted as a set of hierarchical regions stored in a tree structure. Hence, the BPT succeeds in presenting: 1) the decomposition of the image in terms of coherent regions, and 2) the inclusion relations of the regions in the scene. Based on region-merging techniques, the BPT construction is investigated by studying the hyperspectral region models and the associated similarity metrics. Once the BPT is constructed, the fixed tree structure allows implementing efficient and advanced application-dependent techniques on it. The application-dependent processing of BPT is generally implemented through a specific pruning of the tree. In this paper, a pruning strategy is proposed and discussed in a classification context. Experimental results on various hyperspectral data sets demonstrate the interest and the good performances of the BPT representation.

  • Gene expression data classification combining hierarchical representation and efficient feature selection

     Bosio, Mattia; Bellot Pujalte, Pau; Salembier Clairon, Philippe Jean; Oliveras Verges, Albert
    Journal of biological systems
    Date of publication: 2012-12
    Journal article

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    A general framework for microarray data classification is proposed in this paper. It pro- duces precise and reliable classifiers through a two-step approach. At first, the original feature set is enhanced by a new set of features called metagenes. These new features are obtained through a hierarchical clustering process on the original data. Two different metagene generation rules have been analyzed, called Treelets clustering and Euclidean clustering. Metagenes creation is attractive for several reasons: first, they can improve the classification since they broaden the available feature space and capture the com- mon behavior of similar genes reducing the residual measurement noise. Furthermore, by analyzing some of the chosen metagenes for classification with gene set enrichment analysis algorithms, it is shown how metagenes can summarize the behavior of func- tionally related probe sets. Additionally, metagenes can point out, still undocumented, highly discriminant probe sets numerically related to other probes endowed with prior biological information in order to contribute to the knowledge discovery process. The second step of the framework is the feature selection which applies the Improved Sequential Floating Forward Selection algorithm (IFFS) to properly choose a subset from the available feature set for classification composed of genes and metagenes. Considering the microarray sample scarcity problem, besides the classical error rate, a reliability measure is introduced to improve the feature selection process. Different scoring schemes are studied to choose the best one using both error rate and reliability. The Linear Discriminant Analysis classifier (LDA) has been used throughout this work, due to its good characteristics, but the proposed framework can be used with almost any classifier. The potential of the proposed framework has been evaluated analyzing all the publicly available datasets offered by the Micro Array Quality Control Study, phase II (MAQC). The comparative results showed that the proposed framework can compete with a wide variety of state of the art alternatives and it can obtain the best mean performance if a particular setup is chosen. A Monte Carlo simulation confirmed that the proposed framework obtains stable and repeatable results.

  • Distance-based measures of association with applications in relating hyperspectral images

     Cuadras Avellana, Carles M.; Valero Valbuena, Silvia; Cuadras, Daniel; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    Communications in statistics. Simulation and computation
    Date of publication: 2012-07-01
    Journal article

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • Filtering and segmentation of polarimetric SAR data based on binary partition trees

     Alonso Gonzalez, Alberto; López Martínez, Carlos; Salembier Clairon, Philippe Jean
    IEEE transactions on geoscience and remote sensing
    Date of publication: 2012-02
    Journal article

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    In this paper,we propose the use of binary partition trees (BPT) to introduce a novel region-based and multi-scale polarimetric SAR (PolSAR) data representation. The BPT structure represents homogeneous regions in the data at different detail levels. The construction process of the BPT is based, firstly, on a region model able to represent the homogeneous areas, and, secondly, on a dissimilarity measure in order to identify similar areas and define the merging sequence. Depending on the final application, a BPT pruning strategy needs to be introduced. In this paper, we focus on the application of BPT PolSAR data representation for speckle noise filtering and data segmentation on the basis of the Gaussian hypothesis, where the average covariance or coherency matrices are considered as a region model. We introduce and quantitatively analyze different dissimilarity measures. In this case, and with the objective to be sensitive to the complete polarimetric information under the Gaussian hypothesis, dissimilarity measures considering the complete covariance or coherency matrices are employed.When confronted to PolSAR speckle filtering, two pruning strategies are detailed and evaluated. As presented, the BPT PolSAR speckle filter defined filters data according to the complete polarimetric information. As shown, this novel filtering approach is able to achieve very strong filtering while preserving the spatial resolution and the polarimetric information. Finally, the BPT representation structure is employed for high spatial resolution image segmentation applied to coastline detection. The analyses detailed in this work are based on simulated, as well as on real PolSAR data acquired by the ESAR system of DLR and the RADARSAT-2 system.

  • Neighborhood filters and the recovery of 3D information

     Digne, Julie; Dimiccoli, Mariella; Salembier Clairon, Philippe Jean; Sabater, Neus
    Date of publication: 2011
    Book chapter

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • Binary partition tree as a polarimetric SAR data representation in the space-time domain

     Alonso Gonzalez, Alberto; López Martínez, Carlos; Salembier Clairon, Philippe Jean
    IEEE International Geoscience and Remote Sensing Symposium
    Presentation's date: 2011-07
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    The aim of this paper is to present a Polarimetric Synthetic Aperture Radar data processing technique on the space-time domain. This approach is based on a Binary Partition Tree (BPT), which is a region-based and multi-scale data representation. Results with series of RADARSAT-2 real data are analyzed from the point of view of speckle filtering and change detection applications, to illustrate the capabilities to detect and preserve spatial and temporal contours.

  • Hierarchical analysis of remote sensing data: morphological attribute profiles and binary partition trees

     Benediktsson, Joan A; Bruzzone, Lorenzo; Chanussot, Jocelyn; Dalla Mura, Mauro; Salembier Clairon, Philippe Jean; Valero, Silvia
    International Symposium on Mathematical Morphology
    Presentation's date: 2011-07-07
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    The new generation of very high resolution sensors in airborne or satellite remote sensing open the door to countless new applications with a high societal impact. In order to bridge the gap between the potential offered by these new sensors and the needs of the end-users to actually face tomorrows challenges, advanced image processing methods need to be designed. In this paper we discuss two of the most promising strategies aiming at a hierarchical description and analysis of remote sensing data, namely the Extended Attribute Profiles (EAP) and the Binary Partition Trees (BPT). The EAP computes for each pixel a vector of attributes providing a local multiscale representation of the information and hence leading to a fine description of the local structures of the image. Using different attributes allows to address different contexts or applications. The BPTs provide a complete hierarchical description of the image, from the pixels (the leaves) to larger regions as the merging process goes on. The pruning of the tree provides a partition of the image and can address various goals (segmentation, object extraction, classification). The EAP and BPT approaches are used in experiments and the obtained results demonstrate their importance.

  • Improved binary partition tree construction for hyperspectral images: application to object detection

     Valero, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn; Cuadres, Carles
    IEEE International Geoscience and Remote Sensing Symposium
    Presentation's date: 2011-07-27
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    This paper discusses hierarchical region-based representation using Binary Partition Tree in the framework of hyperspectral data. Based on region merging techniques, this region-based representation reduces the number of elementary primitives compared to the pixel based representation and allows a more robust filtering, segmentation, classification or information retrieval. The work presented here proposes a strategy for merging hyperspectral regions using a new association measure depending on canonical correlations relating principal coordinates. To demonstrate an example of BPT usefulness, a pruning strategy aiming at object detection is discussed. Experimental results demonstrate the good performances of BPT.

  • Hyperspectral image segmentation using binary partition trees

     Valero, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    IEEE International Conference on Image Processing
    Presentation's date: 2011-09-12
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    The work presented here proposes a new Binary Partition Tree pruning strategy aimed at the segmentation of hyperspectral images. The BPT is a region-based representation of images that involves a reduced number of elementary primitives and therefore allows to design a robust and efficient segmentation algorithm. Here, the regions contained in the BPT branches are studied by recursive spectral graph partitioning. The goal is to remove subtrees composed of nodes which are considered to be similar. To this end, affinity matrices on the tree branches are computed using a new distance-based measure depending on canonical correlations relating principal coordinates. Experimental results have demonstrated the good performances of BPT construction and pruning.

  • Feature set enhancement via hierarchical clustering for microarray classification

     Bosio, Mattia; Bellot Pujalte, Pau; Salembier Clairon, Philippe Jean; Oliveras Verges, Albert
    IEEE International Workshop on Genomic Signal Processing and Statistics
    Presentation's date: 2011
    Presentation of work at congresses

     Share Reference managers Reference managers Open in new window

  • Occlusion-based depth ordering on monocular images with binary partition tree

     Palou Visa, Guillem; Salembier Clairon, Philippe Jean
    IEEE International Conference on Acoustics, Speech and Signal Processing
    Presentation's date: 2011-05
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    This paper proposes a system to relate objects in an image using occlusion cues and arrange them according to depth. The system does not rely on any a priori knowledge of the scene structure and focuses on detecting specific points, such as T-junctions, to infer the depth relationships between objects in the scene. The system makes extensive use of the Binary Partition Tree (BPT) as the segmentation tool jointly with a new approach for T-junction estimation. Following a bottom-up strategy, regions (initially individual pixels) are iteratively merged until only one region is left. At each merging step, the system estimates the probability of observing a T-junction which is a cue of occlusion when three regions meet. When the BPT is constructed and the pruning is performed, this information is used for depth ordering. Although the proposed system only relies on one low-level depth cue and does not involve any learning process, it shows similar performances than the state of the art.

  • Arbre de partition binaire: un nouvel outil pour la représentation hiérarchique et l¿analyse des images hyperspectrales

     Valero, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    Colloque sur le Traitement du Signal et des Images
    Presentation's date: 2011-09-05
    Presentation of work at congresses

    Read the abstract Read the abstract View View Open in new window  Share Reference managers Reference managers Open in new window

    This problem discusses here is the hierarchical representation and processing of the hyperspectral imaging. In this framework, Binary Partition Trees (BPTs) are proposed as new hierarchical region-based representation. Based on region merging techniques, the work presented here proposes a strategy for merging hyperspectral regions using a new association measure depending on canonical correlations relating principal coordinates. Once is BPT constructed, this representation can be used for many applications including ltering, segmentation and classi cation.To demonstrate an example of BPT usefulness, a pruning strategy aiming at object detection is discussed. Experimental results demonstrate the good performances of BPT.

  • PolSAR speckle filtering and segmentation based on binary partition tree representation

     Alonso Gonzalez, Alberto; López Martínez, Carlos; Salembier Clairon, Philippe Jean
    International Workshop on Science and Applications of SAR Polarimetry and Polarimetric Interferometry
    Presentation's date: 2011-01
    Presentation of work at congresses

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • HYPERSPECTRAL IMAGE REPRESENTATION AND PROCESSING USING BINARY PARTITION TREES

     Valero Valbuena, Silvia
    Defense's date: 2011-12-09
    Department of Signal Theory and Communications, Universitat Politècnica de Catalunya
    Theses

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • SURFACE RECONSTRUCTION FOR MULTI-VIEW VIDEO  Open access

     Salvador Marcos, Jordi
    Defense's date: 2011-09-23
    Department of Signal Theory and Communications, Universitat Politècnica de Catalunya
    Theses

    Read the abstract Read the abstract Access to the full text Access to the full text Open in new window  Share Reference managers Reference managers Open in new window

    This thesis introduces a methodology for obtaining an alternative representation of video sequences captured by calibrated multi-camera systems in controlled environments with known scene background. This representation consists in a 3D description of the surfaces of foreground objects, which allows for the recovering of part of the 3D information of the original scene lost in the projection process in each camera. The choice of the type of representation and the design of the reconstruction techniques are driven by three requirements that appear in smart rooms or recording studios. In these scenarios, video sequences captured by a multi-camera rig are used both for analysis applications and interactive visualization methods. The requirements are: the reconstruction method must be fast in order to be usable in interactive applications, the surface representation must provide a compression of the multi-view data redundancies and this representation must also provide all the relevant information to be used for analysis applications as well as for free-viewpoint video. Once foreground and background are segregated for each view, the reconstruction process is divided in two stages. The first one obtains a sampling of the foreground surfaces (including orientation and texture), whereas the second provides closed, continuous surfaces from the samples, through interpolation. The sampling process is interpreted as a search for 3D positions that result in feature matchings between different views. This search process can be driven by different mechanisms: an image-based approach, another one based on the deformation of a surface from frame to frame or a statistical sampling approach where samples are searched around the positions of other detected samples, which is the fastest and easiest to parallelize of the three approaches. A meshing algorithm is also presented, which allows for the interpolation of surfaces between samples. Starting by an initial triangle, which connects three points coherently oriented, an iterative expansion of the surface over the complete set of samples takes place. The proposed method presents a very accurate reconstruction and results in a correct topology. Furthermore, it is fast enough to be used interactively. The presented methodology for surface reconstruction permits obtaining a fast, compressed and complete representation of foreground elements in multi-view video, as reflected by the experimental results.

    Aquesta tesi presenta diferents tècniques per a la definiciò d’una metodologia per obtenir una representaciò alternativa de les seqüències de vídeo capturades per sistemes multi-càmera calibrats en entorns controlats, amb fons de l’escena conegut. Com el títol de la tesi suggereix, aquesta representació consisteix en una descripció tridimensional de les superfícies dels objectes de primer pla. Aquesta aproximació per la representació de les dades multi-vista permet recuperar part de la informació tridimensional de l’escena original perduda en el procés de projecció que fa cada càmera. L’elecció del tipus de representació i el disseny de les tècniques per la reconstrucció de l’escena responen a tres requeriments que apareixen en entorns controlats del tipus smart room o estudis de gravació, en què les seqüències capturades pel sistema multi-càmera són utilitzades tant per aplicacions d’anàlisi com per diferents mètodes de visualització interactius. El primer requeriment és que el mètode de reconstrucció ha de ser ràpid, per tal de poder-ho utilitzar en aplicacions interactives. El segon és que la representació de les superfícies sigui eficient, de manera que en resulti una compressió de les dades multi-vista. El tercer requeriment és que aquesta representació sigui efectiva, és a dir, que pugui ser utilitzada en aplicacions d’anàlisi, així com per visualitació. Un cop separats els continguts de primer pla i de fons de cada vista –possible en entorns controlats amb fons conegut–, l’estratègia que es segueix en el desenvolupament de la tesi és la de dividir el procés de reconstrucció en dues etapes. La primera consisteix en obtenir un mostreig de les superfícies (incloent orientació i textura). La segona etapa proporciona superfícies tancades, contínues, a partir del conjunt de mostres, mitjançant un procés d’interpolació. El resultat de la primera etapa és un conjunt de punts orientats a l’espai 3D que representen localment la posició, orientació i textura de les superfícies visibles pel conjunt de càmeres. El procés de mostreig s’interpreta com un procés de cerca de posicions 3D que resulten en correspondències de característiques de la imatge entre diferents vistes. Aquest procés de cerca pot ser conduït mitjançant diferents mecanismes, els quals es presenten a la primera part d’aquesta tesi. La primera proposta és fer servir un mètode basat en les imatges que busca mostres de superfície al llarg de la semi-recta que comença al centre de projeccions de cada càmera i passa per un determinat punt de la imatge corresponent. Aquest mètode s’adapta correctament al cas de voler explotar foto-consistència en un escenari estàtic i presenta caracterìstiques favorables per la seva utilizació en GPUs–desitjable–, però no està orientat a explotar les redundàncies temporals existentsen seqüències multi-vista ni proporciona superfícies tancades. El segon mètode efectua la cerca a partir d’una superfície inicial mostrejada que tanca l’espai on es troben els objectes a reconstruir. La cerca en direcció inversa a les normals –apuntant a l’interior– permet obtenir superfícies tancades amb un algorisme que explota la correlació temporal de l’escena per a l’evolució de reconstruccions 3D successives al llarg del temps. Un inconvenient d’aquest mètode és el conjunt d’operacions topològiques sobre la superfície inicial, que en general no són aplicables eficientment en GPUs. La tercera estratègia de mostreig està orientada a la paral·lelització –GPU– i l’explotació de correlacions temporals i espacials en la cerca de mostres de superfície. Definint un espai inicial de cerca que inclou els objectes a reconstruir, es busquen aleatòriament unes quantes mostres llavor sobre la superfície dels objectes. A continuació, es continuen buscant noves mostres de superfície al voltant de cada llavor –procés d’expansió– fins que s’aconsegueix una densitat suficient. Per tal de millorar l’eficiència de la cerca inicial de llavors, es proposa reduir l’espai de cerca, explotant d’una banda correlacions temporals en seqüències multi-vista i de l’altra aplicant multi-resolució. A continuació es procedeix amb l’expansió, que explota la correlació espacial en la distribució de les mostres de superfície. A la segona part de la tesi es presenta un algorisme de mallat que permet interpolar la superfície entre les mostres. A partir d’un triangle inicial, que connecta tres punts coherentment orientats, es procedeix a una expansió iterativa de la superfície sobre el conjunt complet de mostres. En relació amb l’estat de l’art, el mètode proposat presenta una reconstrucció molt precisa (no modifica la posició de les mostres) i resulta en una topologia correcta. A més, és prou ràpid com per ser utilitzable en aplicacions interactives, a diferència de la majoria de mètodes disponibles. Els resultats finals, aplicant ambdues etapes –mostreig i interpolació–, demostren la validesa de la proposta. Les dades experimentals mostren com la metodologia presentada permet obtenir una representació ràpida, eficient –compressió– i efectiva –completa– dels elements de primer pla de l’escena.

  • Post-processing methods for ocean monitoring in sar images  Open access  awarded activity

     Tello Alonso, Marivi
    Defense's date: 2011-02-07
    Department of Signal Theory and Communications, Universitat Politècnica de Catalunya
    Theses

    Read the abstract Read the abstract Access to the full text Access to the full text Open in new window  Share Reference managers Reference managers Open in new window

    In the recent years, a wide interest has been focused on research and development for the operational use of satellite remote sensing tools for Earth observation. Among different types of sensors, Synthetic Aperture Radars (SAR) offer distinctive characteristics, essential for several applications. The observation capability of SAR sensors is independent of the day – night cycle, of the presence of clouds and of the weather conditions in general. Nevertheless, due to the processing involved in the synthesis of SAR images, automatic interpretation of SAR data is awkward. For an extended operational exploitation of SAR products, the development of specific unsupervised techniques for the post-processing of SAR images is necessary. After analyzing the particularities of SAR images, focusing in particular on the elements making their interpretation difficult, this dissertation proposes a set of post-processing techniques for oceanic SAR images, based on time – frequency methods and in particular on the wavelet theory. First, a multiscale algorithm for automatic spot detection in a noisy background has been developed. It is based on the pointwise combination of wavelet coefficients of different bands at the same scale. This technique has been applied to automatic vessel detection and, more specifically, to difficult situations of detection: small fishing ships with low reflectivity. Its efficiency has been compared to other existing algorithms. After that, a method based on the pointwise combination of wavelet subbands at different scales has been proposed. It has been applied to the robust detection of frontiers and linear features. This technique has been employed for the unsupervised detection and monitoring of the coastline in SAR images. Its robustness has been proven through test on a large set of images showing diverse characteristics. Then, the issue of texture analysis in oceanic SAR images has been addressed. An algorithm for the estimation of the local regularity has been proposed, based on the quantification of the wavelet spectrum evolution through scales. A measure of the local fractality is derived from the local regularity. This technique is applied to the detection of oil spills in the ocean surface in SAR satellite based images. The contributions of this dissertation range in two directions. On the one hand, in the direction of multiscale signal processing and, on the other hand, in the direction of automatic interpretation of SAR images. For the multiscale signal processing, a different way of understanding and applying time – scale decompositions (or, equivalently, time - frequency) is proposed. In order to take more efficiently into account the information content in the projected space, the methods proposed carry out the analysis directly in the transformed domain. For the unsupervised interpretation of SAR images, the suitability of the multiscale framework has been justified. Inscribed in this multiscale theory, the novel techniques proposed in this dissertation are simple, flexible, robust and self-contained.

    En els darrers anys, la investigació per a la utilització operacional d’eines de teledetecció satel·litàries per a aplicacions d’observació de la Terra s’està desenvolupant ràpidament. Entre els diferents tipus de sensors satel·litaris existents, els radars d’obertura sintètica o SARs ofereixen unes característiques distintives, que els fan molt interessants per a diferents aplicacions. Tenen capacitat d’observació tant de dia com de nit i no es veuen afectats per la presència de núvols ni per condicions meteorològiques adverses. No obstant, degut al mètode d’adquisició de les imatges SAR, la seva interpretació automàtica és difícil. Per a una explotació operacional i estesa de les dades SAR, és necessari desenvolupar tècniques automàtiques i específiques de post-processament de les imatges SAR. Després d’analitzar les característiques de les imatges SAR que fan que el seu processament automàtic sigui difícil, aquesta tesi proposa un seguit de tècniques de post-processament per a imatges SAR oceàniques basades en mètodes temps – freqüència. En primer lloc, s’ha desenvolupat, mitjançant combinació punt a punt de bandes a una mateixa escala del domini wavelet, un algoritme multiescalar per a la detecció automàtica de spots en entorns sorollosos. Aquest mètode s’ha aplicat a la detecció automàtica de vaixells i, en particular, als casos difícils de detecció: vaixells petits i amb reflectivitat feble. El seu rendiment s’ha comparat amb algoritmes operacionals. En segon lloc, es proposa un mètode basat en la combinació punt a punt de bandes a diferents escales del domini wavelet, per a la detecció robusta de fronteres i línies. Aquesta tècnica s’ha provat per a la detecció automàtica i el seguiment de la línia de costa en imatges SAR. S’ha provat la seva robustesa. Per últim, s’ha plantejat el problema de l’anàlisi de textures a les imatges SAR oceàniques. S’ha proposat un algoritme per a l’estimació de la regularitat local, basat en la quantificació de l’evolució de l’espectre wavelet a través de les escales. A través de la regularitat local, es deriva una mesura de la fractalitat local. Aquesta tècnica s’ha emprat per a la detecció de vessaments de petroli a l’oceà a partir d’imatges SAR. Les contribucions d’aquest treball són a dos nivells. Per una banda, en la direcció del tractament de senyal multiescalar i, per una altra banda, en la direcció de la interpretació automàtica de les imatges SAR. Pel que fa al tractament de senyal multiescalar, es proposa una forma diferent d’utilitzar les descomposicions espai – escala (o, equivalentment, temps - freqüència). Pel tal d’aprofitar de forma més eficaç la informació ressaltada per la projecció, els mètodes proposats realitzen l’anàlisi directament al domini transformat. Pel que fa a la interpretació automàtica de les imatges SAR, s’ha demostrat en primer lloc l’adequació del marc multiescalar que es proposa en aquesta dissertació. Inscrites en aquest marc, les tècniques automàtiques proposades són senzilles, flexibles, robustes i autocontingudes.

  • Fellow of IEEE

     Salembier Clairon, Philippe Jean
    Award or recognition

     Share

  • Best paper award of IEEE International Geoscience and Remote Sensing Symposium

     Valero Valbuena, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    Award or recognition

    View View Open in new window  Share

  • Procesado de vídeo multicámara empleando información de la escena: aplicación a eventos deportivos, interacción visual y 3DTV

     Giro Nieto, Xavier; Oliveras Verges, Albert; Gasull Llampallas, Antoni; Salembier Clairon, Philippe Jean; Marques Acosta, Fernando; Sayrol Clols, Elisa; Pardas Feliu, Montserrat; Morros Rubió, Josep Ramon; Ruiz Hidalgo, Javier; Vilaplana Besler, Veronica; Casas Pla, Josep Ramon
    Participation in a competitive project

     Share

  • Best Paper award of the IEEE International Geosciencie and Remote Sensing Symposium

     Salembier Clairon, Philippe Jean; Valero Valbuena, Silvia
    Award or recognition

    View View Open in new window  Share

  • Access to the full text
    Edge enhancement algorithm based on the wavelet transform for automatic edge detection in SAR images  Open access

     Tello Alonso, Marivi; López Martínez, Carlos; Mallorqui Franquet, Jordi Joan; Salembier Clairon, Philippe Jean
    IEEE transactions on geoscience and remote sensing
    Date of publication: 2011-01-01
    Journal article

    Read the abstract Read the abstract Access to the full text Access to the full text Open in new window  Share Reference managers Reference managers Open in new window

    This paper presents a novel technique for automatic edge enhancement and detection in synthetic aperture radar (SAR) images. The characteristics of SAR images justify the importance of an edge enhancement step prior to edge detection. Therefore, this paper presents a robust and unsupervised edge enhancement algorithm based on a combination of wavelet coefficients at different scales. The performance of the method is first tested on simulated images. Then, in order to complete the automatic detection chain, among the different options for the decision stage, the use of geodesic active contour is proposed. The second part of this paper suggests the extraction of the coastline in SAR images as a particular case of edge detection. Hence, after highlighting its practical interest, the technique that is theoretically presented in the first part of this paper is applied to real scenarios. Finally, the chances of its operational capability are assessed.

  • Access to the full text
    New hyperspectral data representation using binary partition tree  Open access  awarded activity

     Valero Valbuena, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    IEEE International Geoscience and Remote Sensing Symposium
    Presentation's date: 2010-07-29
    Presentation of work at congresses

    Read the abstract Read the abstract Access to the full text Access to the full text Open in new window  Share Reference managers Reference managers Open in new window

    The optimal exploitation of the information provided by hyperspectral images requires the development of advanced image processing tools. This paper introduces a new hierarchical structure representation for such images using binary partition trees (BPT). Based on region merging techniques using statistical measures, this region-based representation reduces the number of elementary primitives and allows a more robust filtering, segmentation, classification or information retrieval. To demonstrate BPT capabilites, we first discuss the construction of BPT in the specific framework of hyperspectral data. We then propose a pruning strategy in order to perform a classification. Labelling each BPT node with SVM classifiers outputs, a pruning decision based on an impurity measure is addressed. Experimental results on two different hyperspectral data sets have demonstrated the good performances of a BPT-based representation

  • Comparison of merging orders and pruning strategies for binary partition tree in hyperspectral data

     Valero Valbuena, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    IEEE International Conference on Image Processing
    Presentation's date: 2010-09-28
    Presentation of work at congresses

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • Access to the full text
    Filtering and segmentation of polarimetric SAR images with binary partition trees  Open access

     Alonso Gonzalez, Alberto; López Martínez, Carlos; Salembier Clairon, Philippe Jean
    IEEE International Geoscience and Remote Sensing Symposium
    Presentation's date: 2010-07
    Presentation of work at congresses

    Read the abstract Read the abstract Access to the full text Access to the full text Open in new window  Share Reference managers Reference managers Open in new window

    A new multi-scale PolSAR data filtering technique, based on a Binary Partition Tree (BPT) representation of the data, is proposed. Different alternatives for the construction and the exploitation of the BPT for filtering and segmentation are presented. Results with simulated and experimental PolSAR data are presented to shown the capabilities of the BPT-filtering strategy to maintain both spatial details and the polarimetric information.

  • Access to the full text
    Improved local pdf estimation in the wavelet domain for generalized lifting  Open access

     Rolon Garrido, Julio Cesar; Salembier Clairon, Philippe Jean
    Picture Coding Symposium
    Presentation's date: 2010-12-08
    Presentation of work at congresses

    Read the abstract Read the abstract Access to the full text Access to the full text Open in new window  Share Reference managers Reference managers Open in new window

    Generalized Lifting (GL) has been studied for lossy image compression in [2,3]. It has been demonstrated that the method leads to a significant reduction of the wavelet coefficients energy and entropy. The definition of the GL relies on an estimation of the pdf of the pixel to encode conditioned to a surrounding context. The objective of this paper is to present an improved method for the estimation of the pdf at the local level. Rather than assuming that the local pdf is monomodal, symmetric, and centered at the central value of the best context match within a neighborhood, as in [3], we follow the idea of self similarity proposed in [1] for denoising, and propose to estimate the pdf using all the causal contexts within a window. Therefore, all the available knowledge about the neighborhood is incorporated. No assumptions about the characteristics of the pdf are made. A generalized lifting operator that minimizes the energy is applied to each context during the encoding process. Experimental results show an important increment in the energy and entropy gains when compared to previous strategies [2,3].

  • Some measures of multivariate association relating two spectral data sets

     Cuadras, C.M.; Valero Valbuena, Silvia; Salembier Clairon, Philippe Jean; Chanussot, Jocelyn
    COMPSTAT International Conference on Computational Statistics
    Presentation's date: 2010-12-22
    Presentation of work at congresses

     Share Reference managers Reference managers Open in new window

  • Compression

     Marcotegui, Beatriz; Salembier Clairon, Philippe Jean
    Date of publication: 2010-09-01
    Book chapter

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • Connected operators based on tree pruning strategies

     Salembier Clairon, Philippe Jean
    Date of publication: 2010-06-01
    Book chapter

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • Compression

     Marcotegui, Beatriz; Salembier Clairon, Philippe Jean
    Date of publication: 2010-06-01
    Book chapter

    View View Open in new window  Share Reference managers Reference managers Open in new window

  • Region-based face detection, segmentation and tracking. Framework definition and application to other objects  Open access

     Vilaplana Besler, Veronica
    Defense's date: 2010-12-17
    Department of Signal Theory and Communications, Universitat Politècnica de Catalunya
    Theses

    Read the abstract Read the abstract Access to the full text Access to the full text Open in new window  Share Reference managers Reference managers Open in new window

    One of the central problems in computer vision is the automatic recognition of object classes. In particular, the detection of the class of human faces is a problem that generates special interest due to the large number of applications that require face detection as a first step. In this thesis we approach the problem of face detection as a joint detection and segmentation problem, in order to precisely localize faces with pixel accurate masks. Even though this is our primary goal, in finding a solution we have tried to create a general framework as independent as possible of the type of object being searched. For that purpose, the technique relies on a hierarchical region-based image model, the Binary Partition Tree, where objects are obtained by the union of regions in an image partition. In this work, this model is optimized for the face detection and segmentation tasks. Different merging and stopping criteria are proposed and compared through a large set of experiments. In the proposed system the intra-class variability of faces is managed within a learning framework. The face class is characterized using a set of descriptors measured on the tree nodes, and a set of one-class classifiers. The system is formed by two strong classifiers. First, a cascade of binary classifiers simplifies the search space, and afterwards, an ensemble of more complex classifiers performs the final classification of the tree nodes. The system is extensively tested on different face data sets, producing accurate segmentations and proving to be quite robust to variations in scale, position, orientation, lighting conditions and background complexity. We show that the technique proposed for faces can be easily adapted to detect other object classes. Since the construction of the image model does not depend on any object class, different objects can be detected and segmented using the appropriate object model on the same image model. New object models can be easily built by selecting and training a suitable set of descriptors and classifiers. Finally, a tracking mechanism is proposed. It combines the efficiency of the mean-shift algorithm with the use of regions to track and segment faces through a video sequence, where both the face and the camera may move. The method is extended to deal with other deformable objects, using a region-based graph-cut method for the final object segmentation at each frame. Experiments show that both mean-shift based trackers produce accurate segmentations even in difficult scenarios such as those with similar object and background colors and fast camera and object movements. Lloc i

    Un dels problemes més importants en l'àrea de visió artificial és el reconeixement automàtic de classes d'objectes. En particular, la detecció de la classe de cares humanes és un problema que genera especial interès degut al gran nombre d'aplicacions que requereixen com a primer pas detectar les cares a l'escena. A aquesta tesis s'analitza el problema de detecció de cares com un problema conjunt de detecció i segmentació, per tal de localitzar de manera precisa les cares a l'escena amb màscares que arribin a precisions d'un píxel. Malgrat l'objectiu principal de la tesi és aquest, en el procés de trobar una solució s'ha intentat crear un marc de treball general i tan independent com fos possible del tipus d'objecte que s'està buscant. Amb aquest propòsit, la tècnica proposada fa ús d'un model jeràrquic d'imatge basat en regions, l'arbre binari de particions (BPT: Binary Partition Tree), en el qual els objectes s'obtenen com a unió de regions que provenen d'una partició de la imatge. En aquest treball, s'ha optimitzat el model per a les tasques de detecció i segmentació de cares. Per això, es proposen diferents criteris de fusió i de parada, els quals es comparen en un conjunt ampli d'experiments. En el sistema proposat, la variabilitat dins de la classe cara s'estudia dins d'un marc de treball d'aprenentatge automàtic. La classe cara es caracteritza fent servir un conjunt de descriptors, que es mesuren en els nodes de l'arbre, així com un conjunt de classificadors d'una única classe. El sistema està format per dos classificadors forts. Primer s'utilitza una cascada de classificadors binaris que realitzen una simplificació de l'espai de cerca i, posteriorment, s'aplica un conjunt de classificadors més complexes que produeixen la classificació final dels nodes de l'arbre. El sistema es testeja de manera exhaustiva sobre diferents bases de dades de cares, sobre les quals s'obtenen segmentacions precises provant així la robustesa del sistema en front a variacions d'escala, posició, orientació, condicions d'il·luminació i complexitat del fons de l'escena. A aquesta tesi es mostra també que la tècnica proposada per cares pot ser fàcilment adaptable a la detecció i segmentació d'altres classes d'objectes. Donat que la construcció del model d'imatge no depèn de la classe d'objecte que es pretén buscar, es pot detectar i segmentar diferents classes d'objectes fent servir, sobre el mateix model d'imatge, el model d'objecte apropiat. Nous models d'objecte poden ser fàcilment construïts mitjançant la selecció i l'entrenament d'un conjunt adient de descriptors i classificadors. Finalment, es proposa un mecanisme de seguiment. Aquest mecanisme combina l'eficiència de l'algorisme mean-shift amb l'ús de regions per fer el seguiment i segmentar les cares al llarg d'una seqüència de vídeo a la qual tant la càmera com la cara es poden moure. Aquest mètode s'estén al cas de seguiment d'altres objectes deformables, utilitzant una versió basada en regions de la tècnica de graph-cut per obtenir la segmentació final de l'objecte a cada imatge. Els experiments realitzats mostren que les dues versions del sistema de seguiment basat en l'algorisme mean-shift produeixen segmentacions acurades, fins i tot en entorns complicats com ara quan l'objecte i el fons de l'escena presenten colors similars o quan es produeix un moviment ràpid, ja sigui de la càmera o de l'objecte.

  • Information theoretical region merging approaches and fusion of hierarchical image segmentation results

     Calderero Patino, Felipe
    Defense's date: 2010-02-12
    Department of Signal Theory and Communications, Universitat Politècnica de Catalunya
    Theses

     Share Reference managers Reference managers Open in new window

  • Generalized lifting for sparse image representation and coding

     Rolon Garrido, Julio Cesar
    Defense's date: 2010-01-25
    Department of Signal Theory and Communications, Universitat Politècnica de Catalunya
    Theses

     Share Reference managers Reference managers Open in new window

  • Adquisición multicámara para Free Viewpoint Video (MC4FVV)

     Pardas Feliu, Montserrat; Giro Nieto, Xavier; Vilaplana Besler, Veronica; Ruiz Hidalgo, Javier; Morros Rubió, Josep Ramon; Salembier Clairon, Philippe Jean; Marques Acosta, Fernando; Gasull Llampallas, Antoni; Oliveras Verges, Albert; Sayrol Clols, Elisa; Casas Pla, Josep Ramon
    Participation in a competitive project

     Share