2015

Journal articles

titre
Geometrical cues in visual saliency models for active object recognition in egocentric videos
auteur
Vincent Buso, Jenny Benois-Pineau, Jean-Philippe Domenger
article
Multimedia Tools and Applications, Springer Verlag, 2015, Multimedia Tools and Applications, 74 (22), pp.10077 - 10095. <http://link.springer.com/article/10.1007/s11042-015-2803-2>. <10.1007/s11042-015-2803-2>
identifiant
hal-01436937
resume
In the problem of “human sensing”, videos recorded with wearable cameras give an “egocentric” view of the world, capturing details of human activities. In this paper we continue research on visual saliency for such kind of content with the goal of “active” objects recognition in egocentric videos. In particular, a geometrical cue is considered in case when the central-bias hypothesis does not hold. The proposed visual saliency models are trained based on eye fixations of observers and incorporated into spatio-temporal saliency models. The proposed models have been compared to state of the art visual saliency models using a metric based on target object recognition performances. The results are promising:they highlight the necessity of a non-centered geometric saliency cue.
DOI
DOI : 10.1007/s11042-015-2803-2
Accès au bibtex
BibTex

2013

Journal articles

titre
Quality evaluation of degraded document images for binarization result prediction
auteur
Vincent Rabeux, Nicholas Journet, Anne Vialard, Jean-Philippe Domenger
article
International Journal on Document Analysis and Recognition (IJDAR), 2013, pp.1--13. <10.1007/s10032-013-0211-6>
identifiant
hal-00862234
resume
This article proposes an approach to predict the result of binarization algorithms on a given docu- ment image according to its state of degradation. In- deed, historical documents suffer from different types of degradation which result in binarization errors. We intend to characterize the degradation of a document image by using different features based on the inten- sity, quantity and location of the degradation. These features allow us to build prediction models of bina- rization algorithms that are very accurate according to R2 values and p-values. The prediction models are used to select the best binarization algorithm for a given doc- ument image. Obviously, this image-by-image strategy improves the binarization of the entire dataset.
DOI
DOI : 10.1007/s10032-013-0211-6
Accès au texte intégral et bibtex
https://hal.archives-ouvertes.fr/hal-00862234/file/ijdar.pdf BibTex

2012

Journal articles

titre
Terahertz radiation for tomographic inspection
auteur
Benoît Recur, Jean Paul Guillet, Lena Bassel, Carole Fragnol, Inka Manek-Hönninger, Jean-Christophe Delagnes, William Benharbone, Pascal Desbarats, Jean-Philippe Domenger, Patrick Mounaix
article
Optical Engineering, SPIE, 2012, 51 (9), pp.091609-1 -- 091609-7. <10.1117/1.OE.51.9.091609>
identifiant
hal-00699947
resume
Three-dimensional (3-D) terahertz computed tomography has already been performed with three different reconstruction methods (standard back-projection algorithm and two iterative analyses) to reconstruct 3-D objects. A Gaussian beam model is developed according to the physical properties of terahertz waves such as the energy distribution within the propagation path. This model is included as a new convolution filter into the tomographic reconstruction methods in order to analyze the impact of a such effect and then to enhance quality and accuracy of the resulting images. We demonstrate the improvements of the optimized reconstructions for applied 3-D terahertz tomography.
DOI
DOI : 10.1117/1.OE.51.9.091609
Accès au texte intégral et bibtex
https://hal.archives-ouvertes.fr/hal-00699947/file/091609_1_1337638728_1.pdf BibTex
titre
Propagation Beam Consideration for 3D THz Computed Tomography
auteur
Benoît Recur, Jean-Paul Guillet, Inka Manek-Hönninger, Jean-Christophe Delagnes, William Benharbone, Pascal Desbarats, Jean-Philippe Domenger, Lionel Canioni, Patrick Mounaix
article
Optics Express, Optical Society of America, 2012, 20 (5), pp.5817-5829. <10.1364/OE.20.005817>
identifiant
hal-00672679
resume
In this paper, a model of the beam propagation is developed according to the physical properties of THz waves used in THz computed tomography (CT) scan imaging. This model is first included in an acquisition simulator to observe and estimate the impact of the Gaussian beam intensity profile on the projection sets. Second, the model is introduced in several inversion methods as a convolution filter to perform efficient tomographic reconstructions of simulated and real acquired objects. Results obtained with three reconstruction methods (BFP, SART and OSEM) are compared to the techniques proposed in this paper. We focus our discussion on the efficiency of optimized algorithms to increase the overall quality and accuracy of the reconstructions.
DOI
DOI : 10.1364/OE.20.005817
Accès au texte intégral et bibtex
https://hal.archives-ouvertes.fr/hal-00672679/file/main.pdf BibTex

2011

Journal articles

titre
Segmentation-based multi-class semantic object detection
auteur
Remi Vieux, Jenny Benois-Pineau, Jean-Philippe Domenger, Achille Braquelaire
article
Multimedia Tools and Applications, Springer Verlag, 2011, pp.1 - 22. <10.1007/s11042-010-0611-2>
identifiant
hal-00572863
resume
In this paper we study the problem of the detection of semantic objects from known categories in images. Unlike existing techniques which operate at the pixel or at a patch level for recognition, we propose to rely on the categorization of image segments. Recent work has highlighted that image segments provide a sound support for visual object class recognition. In this work, we use image segments as primitives to extract robust features and train detection models for a predefined set of categories. Several segmentation algorithms are benchmarked and their performances for segment recognition are compared. We then propose two methods for enhancing the segments classification, one based on the fusion of the classification results obtained with the different segmentations, the other one based on the optimization of the global labelling by correcting local ambiguities between neighbor segments. We use as a benchmark the Microsoft MSRC-21 image database and show that our method competes with the current state-of-the-art.
DOI
DOI : 10.1007/s11042-010-0611-2
Accès au texte intégral et bibtex
https://hal.archives-ouvertes.fr/hal-00572863/file/template.pdf BibTex
titre
Local object-based super-resolution mosaicing from low-resolution video
auteur
Petra Krämer, Jenny Benois-Pineau, Jean-Philippe Domenger
article
Signal Processing, Elsevier, 2011, 91 (8), pp.1771-1780
identifiant
hal-00714650
resume
Many efficient super-resolution methods have been presented in the past for improving resolution of images degraded by global blurs. Unfortunately, in video, more complex situations can be observed where local blurs appear in each frame which are mainly caused by object motions. To address this problem, we propose in this article a local super-resolution method which allows the restoration of such local blurs. Moreover, the motion of objects in video sequences may be very complex and particularly in very low-resolution sequences it is difficult to estimate their motion exactly enough to superimpose them for super-resolution. To this end, we present a generic method: An interpolation method is proposed to improve the resolution of moving objects and we derive from this a super-resolution method for the scene background.
Accès au texte intégral et bibtex
https://hal.archives-ouvertes.fr/hal-00714650/file/elsarticle-template-num.pdf BibTex

2010

Journal articles

titre
Noise Robust Mojette Reconstructions for Missing Wedge Effect Attenuation
auteur
Benoît Recur, Pascal Desbarats, Jean-Philippe Domenger
article
Journal of Multimedia Processing and Technologies, 2010, 1 (4), pp.208-227
identifiant
hal-00627493
resume
The Mojette transform is a tomographic reconstruction method based on a discrete and finite interpretation of the Radon theorem. Since the Mojette acquisition follows the discrete image geometry, this method resolves the well-known irregular sampling problem. A specific algorithm called Corner Based Inversion (CBI) is proposed to reconstruct without any error an image from its projections even if the angular coverage is not sufficient (missing wedge). However, this reconstruction is noise sensitive and reconstruction from corrupted data fails. In this paper, we develop new noise robust CBI algorithms and we apply them both on discrete Mojette acquisitions and on usual Radon acquisitions. Reconstruction results are discussed to highlight the efficiency of these algorithms for usual tomography and perspectives are proposed to reduce the missing wedge effect.
Accès au texte intégral et bibtex
https://hal.archives-ouvertes.fr/hal-00627493/file/JMPT.pdf BibTex
titre
Scalable Object-based Video Retrieval in HD Video DataBases
auteur
Claire Morand, Jenny Benois-Pineau, Jean-Philippe Domenger, Joaquin Zepeda, Ewa Kijak, Christine Guillemot
article
Eurasip Journal on Signal Processing: Image Communication, Hindawi, 2010, accepted
identifiant
inria-00504257
resume
No abstract
Accès au bibtex
BibTex

2007

Journal articles

titre
Retrieval of objects in video by similarity based on graph matching
auteur
Fanny Chevalier, Jean-Philippe Domenger, Jenny Benois-Pineau, Maylis Delest
article
Pattern Recognition Letters, Elsevier, 2007, 28, pp.939--949
identifiant
hal-00307872
Accès au bibtex
BibTex
titre
A heuristic for the retrieval of objects in video in the framework of the rough indexing paradigm
auteur
Maylis Delest, Fanny Chevalier, Jean-Philippe Domenger
article
Signal Processing : Image Communication, 2007, 22, pp.622--634
identifiant
hal-00354274
resume
In this paper, we tackle the problem of matching of objects in video in the framework of the rough indexing paradigm. In this context, the video data are of very low spatial and temporal resolution because they come from partially decoded MPEG compressed streams. This paradigm enables us to achieve our purpose in near real time due to the faster computation on rough data than on original full spatial and temporal resolution video frames. In this context, segmentation of rough video frames is inaccurate and the region features (texture, color, shape) are not strongly relevant. The structure of the objects must be considered in order to improve the robustness of the matching of regions. The problem of object matching can be expressed in terms of region adjacency graph (RAG) matching. Here, we propose a directed acyclic graph (DAG) matching method based on a heuristic in order to approximate object matching. The RAGs to compare are first transformed into DAGs by orienting edges. Then, we compute some combinatoric metrics on nodes in order to classify them by similarity. At the end, a top-down process on DAGs aims to match similar patterns that exist between the two DAGs. The results are compared with those of a method based on relaxation matching.
Accès au bibtex
BibTex
titre
Super-Resolution Mosaicing from MPEG Compressed Video
auteur
P. Kraemer, O. Hadar, Jenny Benois-Pineau, Jean-Philippe Domenger
article
Signal Processing: Image Communication, Elsevier, 2007, pp.I - 893-6
identifiant
hal-00308210
Accès au bibtex
BibTex

2006

Journal articles

titre
Scene Similarity Measure for Video Content Segmentation in the Framework of a Rough Indexing Paradigm
auteur
P. Kraemer, Jenny Benois-Pineau, Jean-Philippe Domenger
article
International Journal of Intelligent Systems, Wiley, 2006, 21, pp.765-783
identifiant
hal-00308208
Accès au bibtex
BibTex
titre
Efficient drawing and comparison of RNA secondary structure
auteur
David Auber, Maylis Delest, Serge Dulucq, Jean-Philippe Domenger
article
Journal of Graph Algorithms and Applications (JGAA), Brown University, 2006, 10, pp.329-351
identifiant
hal-00307625
Accès au bibtex
BibTex

1998

Journal articles

titre
Representation of segmented images with discrete geometric maps
auteur
Achille Braquelaire, Jean-Philippe Domenger
article
Image and Vision Computing, Elsevier, 1998, 17, pp.715-735
identifiant
hal-01465410
resume
Most models of representation of a segmented image have been designed to perform specific processing on images and are generally illadapted to other processes. Moreover, usually models describe separately the geometry and the topology of the regions of the image. Generally, this dichotomy makes it difficult or costly to design and implement operations involving both topological and geometrical features. In this paper, we give a complete description of a model encoding both geometry and topology of the region’s of a segmented image. This model, based on inter-pixel boundaries associated with topological planar maps, has been successfully used for image segmentation.
Accès au bibtex
BibTex

1997

Journal articles

titre
Geometrical, topological and hierarchical structuring of overlapping 2-d discrete objects.
auteur
Achille Braquelaire, Jean-Philippe Domenger
article
Computers and Graphics, Elsevier, 1997, 21 (5), pp.513-522
identifiant
hal-01464149
resume
A model of scenes composed of 2-D discrete objects is presented. A generalization of the 2 1/2-D positioning features usually implemented in 2 1/2-D environments is given, treating object superimposition. The positioning representation sits on an explicit topological and geometrical encoding of object boundaries. This approach allows structured manipulations of objects (intersection, combination,...) and efficient redisplay of scene, such that every modified pixel is drawn exactly once. Moreover, this model is consistent from a geometrical and topological point of view and allows the construction of a scene from a digital image. Many applications can be considered; graphics editors, paintboxes, image processing environments, etc.
Accès au bibtex
BibTex