Filters








354 Hits in 7.1 sec

Using Topic Concepts for Semantic Video Shots Classification [chapter]

Stéphane Ayache, Georges Quénot, Jérôme Gensel, Shin'ichi Satoh
2006 Lecture Notes in Computer Science  
We have conducted several experiments on the TRECVID'05 collection and show here that several intermediate topic classifiers can bridge parts of the semantic gap and help to detect high-level concepts.  ...  In this paper, we propose to enhance the classification of highlevel concepts using intermediate topic concepts and study various fusion strategies to combine topic concepts with visual features in order  ...  For 'Global' run, we have directly classified high-level concepts using their corresponding global low level features.  ... 
doi:10.1007/11788034_31 fatcat:u2tdrsxx6fbwvda4x2tac5s2lm

A Hybrid Approach to Improving Semantic Extraction of News Video

A.G. Hauptmann, M.-Y. Chen, M. Christel, W.-H. Lin, J. Yang
2007 International Conference on Semantic Computing (ICSC 2007)  
No single approach provides a consistently better result for every concept detection, which suggests that extracting video semantics should exploit multiple resources and techniques rather than a single  ...  Experiments show the value of careful parameter tuning, exploiting multiple feature sets and multilingual linguistic resources, applying text retrieval approaches for image features, and establishing synergy  ...  Acknowledgements This material is based in part on work supported by the National Science Foundation Grant IIS-0205219.  ... 
doi:10.1109/icosc.2007.4338335 fatcat:ykdrvgyaafbcnhu3yphg64pxx4

A Hybrid Approach to Improving Semantic Extraction of News Video

A.G. Hauptmann, M.-Y. Chen, M. Christel, W.-H. Lin, J. Yang
2007 International Conference on Semantic Computing (ICSC 2007)  
No single approach provides a consistently better result for every concept detection, which suggests that extracting video semantics should exploit multiple resources and techniques rather than a single  ...  Experiments show the value of careful parameter tuning, exploiting multiple feature sets and multilingual linguistic resources, applying text retrieval approaches for image features, and establishing synergy  ...  Since so far only very few high-level concepts can machine reliably extracted, the quest for developing better concept classifiers is never ending.  ... 
doi:10.1109/icsc.2007.68 dblp:conf/semco/HauptmannCCLY07 fatcat:nn54qdfcobhrfplxcr7tgouyrm

A Multi-Pronged Approach to Improving Semantic Extraction of News Video

A. G. Hauptmann, M.-Y. Chen, M. Christel, W.-H. Lin, J. Yang
2009 Journal of Signal Processing Systems  
We also describe an approximate parameter estimation algorithm and present results obtained from the TRECVID 2006 data.  ...  We present a discriminative learning framework called Multi-concept Discriminative Random Field (MDRF) for building probabilistic models of video semantic concept detectors by incorporating related concepts  ...  Acknowledgements This material is based in part on work supported by the National Science Foundation under Grant No. IIS-0205219.  ... 
doi:10.1007/s11265-009-0382-z fatcat:etix4nj4efbbtjndybqyvicaya

Semantic Model Vectors for Complex Video Event Recognition

Michele Merler, Bert Huang, Lexing Xie, Gang Hua, Apostol Natsev
2012 IEEE transactions on multimedia  
The semantic model vectors are extracted using a set of discriminative semantic classifiers, each being an ensemble of SVM models trained from thousands of labeled web images, for a total of 280 generic  ...  We hence present an end-to-end video event detection system, which combines semantic model vectors with other static or dynamic visual descriptors, extracted at the frame, segment, or full clip level.  ...  ACKNOWLEDGMENT The authors would like to thank the Associate Editor and the Reviewers for their insightful comments and suggestions, which contributed to improve and make this work complete.  ... 
doi:10.1109/tmm.2011.2168948 fatcat:ml6wb6yefjh3nlv46sibr6ccjm

A Generalized Discriminative Muitiple Instance Learning for Multimedia Semantic Concept Detection

Sheng Gao, Qibin Sun
2006 2006 International Conference on Image Processing  
We evaluate the GD-MIL on the development set of TRECVID 2005 for high-level feature extraction task. The significant improvement is observed using the GD-MIL over the benchmark.  ...  In the paper we present a generalized discriminative multiple instance learning algorithm (GD-MIL) for multimedia semantic concept detection.  ...  We evaluate the GD-MIL on the development set of TRECVID 2005 for high-level feature extraction task. The significant improvement is observed using the GD-MIL.  ... 
doi:10.1109/icip.2006.313036 dblp:conf/icip/GaoS06 fatcat:cpfbxt6v5vct5bbvnmuvx4ahwy

High-level event detection in video exploiting discriminant concepts

Nikolaos Gkalelis, Vasileios Mezaris, Ioannis Kompatsiaris
2011 2011 9th International Workshop on Content-Based Multimedia Indexing (CBMI)  
Subsequently, an improved subclass discriminant analysis method is used to derive a concept subspace for detecting and recognizing high-level events.  ...  In this paper a new approach to video event detection is presented, combining visual concept detection scores with a new dimensionality reduction technique.  ...  one SVM classifier d κ () for each concept h κ .  ... 
doi:10.1109/cbmi.2011.5972525 dblp:conf/cbmi/GkalelisMK11 fatcat:dxkm6yps5jfwnedtqzfas5wlbq

Exploiting Concept Association to Boost Multimedia Semantic Concept Detection

Sheng Gao, Xinglei Zhu, Qibin Sun
2007 2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07  
We evaluate its efficiency on the task of semantic concept detection on the large-scale news video dataset from TRECVID 2005 development set.  ...  We discuss two usages of association strength: 1) applying positive concepts with high association strength for selecting expressive component in the model-based fusion and 2) applying negative concepts  ...  For each type of feature and concept, one SVM classifier (SVM) or linear discriminative function (LDF) classifier is trained.  ... 
doi:10.1109/icassp.2007.366074 dblp:conf/icassp/GaoZS07 fatcat:6tsqvroki5ggjf3s4wuegofxga

Semantic Video Search

A.W.M. Smeulders, J.C. van Gemert, B. Huurnink, D.C. Koelma, O. de Rooij, K.E.A. van de Sande, C.G.M. Snoek, C.J. Veenman, M. Worring
2007 14th International Conference of Image Analysis and Processing - Workshops (ICIAPW 2007)  
We extract image features, on global, regional, and keypoint level, which we combine with various supervised learners.  ...  The MediaMill team participated in two tasks: concept detection and search. For concept detection we use the MediaMill Challenge as experimental platform.  ...  We extract image features on three levels of abstraction, namely: global level, region level, and keypoint level.  ... 
doi:10.1109/iciapw.2007.39 fatcat:lsb6nuqx4nefjb2higadnf47ze

Acoustic super models for large scale video event detection

Robert Mertens, Howard Lei, Luke Gottlieb, Gerald Friedland, Ajay Divakaran
2011 Proceedings of the 2011 joint ACM workshop on Modeling and representing events - J-MRE '11  
This article presents an approach for event classification that enables searching for arbitrary events, including more abstract concepts, in found video collections based on the analysis of the audio track  ...  More importantly, there is a demand for event detectors that go beyond the simple finding of objects but rather detect more abstract concepts, such as "feeding an animal" or a "wedding ceremony".  ...  The results presented in this paper show that learning based on arbitrary low level audio features can be used to detect high level concepts like those given in TRECVid MED 2011.  ... 
doi:10.1145/2072508.2072513 fatcat:csouubzvmrbr3p647rjotuunwe

Sequence Multi-Labeling: A Unified Video Annotation Scheme With Spatial and Temporal Context

Yuanning Li, Yonghong Tian, Ling-Yu Duan, Jingjing Yang, Tiejun Huang, Wen Gao
2010 IEEE transactions on multimedia  
In SVM SML , a joint kernel is employed to model the feature-level and concept-level context relationships (i.e., the dependencies of concepts on the low-level features, spatial and temporal correlations  ...  Extensive experiments on TRECVID'05 and TRECVID'07 datasets have shown that our proposed SVM SML gains superior performance over the state-of-the-art.  ...  However, due to the well-known gap between low-level features and high-level concepts [4] , the concept detectors purely relying on low-level features only could not achieve desirable performance.  ... 
doi:10.1109/tmm.2010.2066960 fatcat:bzfkkyjykfecjnx4cvyp2skia4

Towards optimal bag-of-features for object categorization and semantic video retrieval

Yu-Gang Jiang, Chong-Wah Ngo, Jun Yang
2007 Proceedings of the 6th ACM international conference on Image and video retrieval - CIVR '07  
We also show that the BoF is highly complementary to global features. By incorporating the BoF with color and texture features, an improvement of 50% is reported on TRECVID-2006 dataset.  ...  For the weighting scheme, we propose a novel soft-weighting method to assess the significance of a visual word to an image.  ...  Here we focus our review on feature-level analysis which is related to our latter experimental comparison.  ... 
doi:10.1145/1282280.1282352 dblp:conf/civr/JiangNY07 fatcat:ww26ir6h5ng7nfxcfbnvqilhbu

Image and Video Indexing Using Networks of Operators

Stéphane Ayache, Georges Quénot, Jérôme Gensel
2007 EURASIP Journal on Image and Video Processing  
Concepts are extracted from the raw signal using networks of operators operating on numcepts.  ...  Low-level descriptors can be fused with intermediate of final concepts.  ...  For the "global" run, we have directly classified high-level concepts using their corresponding global low-level features.  ... 
doi:10.1186/1687-5281-2007-056928 fatcat:y7ztdrv6wzdpnfddk2fe4ojrdy

Image and Video Indexing Using Networks of Operators

Stéphane Ayache, Georges Quénot, Jérôme Gensel
2007 EURASIP Journal on Image and Video Processing  
Concepts are extracted from the raw signal using networks of operators operating on numcepts.  ...  Low-level descriptors can be fused with intermediate of final concepts.  ...  For the "global" run, we have directly classified high-level concepts using their corresponding global low-level features.  ... 
doi:10.1155/2007/56928 fatcat:dsuusilwy5btras4mxakp3xh7e

Perplexity-based evidential neural network classifier fusion using mpeg-7 low-level visual features

Rachid Benmokhtar, Benoit Huet
2008 Proceeding of the 1st ACM international conference on Multimedia information retrieval - MIR '08  
The experimental results conducted in the framework of the TRECVid'07 high level features extraction task report the efficiency and the improvement provided by the proposed scheme.  ...  the shot must be addressed, in the goal of bridging the semantic gap between the low level visual feature and the high level semantic concepts.  ...  EXPERIMENTAL RESULTS The high level concept detection experiments presented in this paper are conducted on the TRECVid 2007 dataset [16] containing science news, news reports, documentaries, educational  ... 
doi:10.1145/1460096.1460151 dblp:conf/mir/BenmokhtarH08 fatcat:5qrdawprinf4hct5vzoj4vykge
« Previous Showing results 1 — 15 out of 354 results