logo EDITE Matthieu CORD
Identité
Matthieu CORD
État académique
Thèse soutenue
Titulaire d'une HDR (ou équivalent) 2004-12-16
Laboratoire: personnel permanent
Direction de thèses (depuis 2007)
5.8
Voisinage
Ellipse bleue: doctorant, ellipse jaune: docteur, rectangle vert: permanent, rectangle jaune: HDR. Trait vert: encadrant de thèse, trait bleu: directeur de thèse, pointillé: jury d'évaluation à mi-parcours ou jury de thèse.
Productions scientifiques
978 1 4577 1302 6
BOSSA: extended BoW formalism for image classification
In image classification, the most powerful statistical learning approaches are based on the Bag-of-Words paradigm. In this article, we propose an extension of this formalism. Considering the Bag-of-Features, dictionary coding and pooling steps, we propose to focus on the pooling step. Instead of using the classical sum or max pooling strategies, we introduced a density function-based pooling strategy. This flexible formalism allows us to better represent the links between dictionary codewords and local descriptors in the resulting image signature. We evaluate our approach in two very challenging tasks of video and image classification, involving very high level semantic categories with large and nuanced visual diversity.
IEEE International Conference on Image Processing (ICIP) 2011
edite:1332792355180
Efficient Bag-of-Feature kernel representation for image similarity search
IEEE International Conference on Image Processing (ICIP) 2011
edite:1332792373227
HMAX-S: DEEP SCALE REPRESENTATION FOR BIOLOGICALLY INSPIRED IMAGE
This paper presents an improvement on a biologically inspired net-
IEEE International Conference on Image Processing 2011
edite:1332792381260
Learning Invariant Color Features with Sparse Topographic RBM
IEEE International Conference on Image Processing (ICIP) 2011
edite:1332792419411
Text Detection and Recognition in Urban Scenes
2011
edite:1332792434500
An Application of Swarm Intelligence to Distributed Image Retrieval
2010
edite:1332792435504
An efficient System for combining complementary kernels in complex visual categorization tasks
2010
edite:1332792443535
Biasing Restricted Boltzmann Machines to Manipulate Latent Selectivity and Sparsity
NIPS 2010 Workshop on Deep Learning and Unsupervised Feature Learning 2010
edite:1332792500825
SnooperText: A Multiresolution System for Text Detection in Complex Visual Scenes
2010
edite:13327925441105
Geometric Consistency Checking for Local-Descriptor Based Document Retrieval
ACM symposium on Document engineering 2009
edite:13327925551209
Optimization on active learning strategy for object category retrieval
IEEE International Conference on Image Processing (ICIP) 2009
edite:13327925621271
Spatio-temporal tube kernel for actor retrieval
IEEE International Conference on Image Processing (ICIP) 2009
edite:13327925621287
Text segmentation in natural scenes using toggle-mapping
IEEE International Conference on Image Processing (ICIP) 2009
oai:hal.archives-ouvertes.fr:hal-00536604
Indexing Personal Image Collections: A Flexible, Scalable Solution
The growth of personal image collections has boosted the creation of many applications, many of which depend on the existence of fast schemes to match similar image descriptors. In this paper we present multicurves, a new indexing method for multimedia descriptors, able to handle high dimensionalities (100 dimensions and over) and large databases (millions of descriptors). The technique allows a fast implementation of approximate kNN search, and deals easily with data updating (insertions and deletions). The index is based on the simultaneous use of several moderate-dimensional space-filling curves. The combined effect of having more than one curve, and reducing the dimensionality of each individual curve allows to overcome undesirable boundary effects. In empirical evaluations, the method compares favorably with state-of-the-art methods, especially when the constraints of secondary storage are considered.
IEEE Transactions on Consumer Electronicspeer-reviewed article 2010-08
oai:hal.archives-ouvertes.fr:hal-00625414
SNOOPERTRACK: TEXT DETECTION AND TRACKING FOR OUTDOOR VIDEOS
In this work we introduced SnooperTrack, an algorithm for the automatic detection and tracking of text objects -- such as store names, traffic signs, license plates, and advertisements -- in videos of outdoor scenes. The purpose is to improve the performances of text detection process in still images by taking advantage of the temporal coherence in videos. We first propose an efficient tracking algorithm using particle filtering framework with original region descriptors. The second contribution is our strategy to merge tracked regions and new detections. We also propose an improved version of our previously published text detection algorithm in still images. Tests indicate that SnooperTrack is fast, robust, enable false positive suppression, and achieved great performances in complex videos of outdoor scenes.
IEEE International Conference on Image Processing (ICIP)proceeding, seminar, workshop without peer review 2011-09-11