Filters








6,982 Hits in 5.1 sec

Cross-domain video concept detection using adaptive svms

Jun Yang, Rong Yan, Alexander G. Hauptmann
2007 Proceedings of the 15th international conference on Multimedia - MULTIMEDIA '07  
One example is cross-domain video concept detection which aims to adapt concept classifiers across various video domains.  ...  The proposed method outperforms several baseline and competing methods in terms of classification accuracy and efficiency in cross-domain concept detection in the TRECVID corpus.  ...  CROSS-DOMAIN VIDEO CONCEPT DETECTION The Task and Collection Video concept detection is to automatically classify video shots by the presence or absence of certain semantic concepts, such as Studio,  ... 
doi:10.1145/1291233.1291276 dblp:conf/mm/YangYH07 fatcat:ou3c5ewvbzcunnf23fmji2gzs4

Video concept detection by learning from web images: A case study on cross domain learning

Shiai Zhu, Ting Yao, Chong-Wah Ngo
2013 2013 IEEE International Conference on Multimedia and Expo Workshops (ICMEW)  
Therefore, in this paper, we consider exploring weakly tagged Web images to shed some light on video concept detection.  ...  Particularly, two sets of Web images downloaded from Flickr are utilized as training data for concept detection on two real-world large-scale video datasets released by TRECVID.  ...  Then, cross-domain adaptive model is trained based on the Tradaboost [13] method using the training data both in source and target domains.  ... 
doi:10.1109/icmew.2013.6618377 dblp:conf/icmcs/ZhuYN13 fatcat:xyv6ae7c2bcwtowokzqj2kgmle

Laplacian adaptive context-based SVM for video concept detection

Wei Jiang, Alexander Loui
2011 Proceedings of the 3rd ACM SIGMM international workshop on Social media - WSM '11  
To accommodate these issues, we propose a Laplacian Adaptive Context-based SVM (LAC-SVM) algorithm that jointly uses four techniques to enhance classification: cross-domain learning that adapts previous  ...  Specifically, LAC-SVM adaptively applies concept classifiers and concept affinity relations computed from a source domain to classify data in the target domain, and at the same time, incrementally updates  ...  CONCLUSION We propose an LAC-SVM approach to improve concept detection by jointly using cross-domain, semi-supervised, multi-concept, and active learning.  ... 
doi:10.1145/2072609.2072615 fatcat:p4cwv6wnxfdz5ir3xgzvpnnsxi

DFKI-IUPR participation in TRECVID'09 High-level Feature Extraction Task

Damian Borth, Markus Koch, Adrian Ulges, Thomas M. Breuel
2009 TREC Video Retrieval Evaluation  
Such an adaptation would be not suitable for SVMs, where cross-domain learning is performed differently and therefore is also computational more expensive.  ...  In terms of cross-domain learning terminology, the YOUTUBE data is our source domain, whereas TRECVID defines our target domain to which we want to adapt.  ... 
dblp:conf/trecvid/BorthKUB09 fatcat:cz2qy2itujh43mowmrivnjrk4y

A pseudo relevance feedback based cross domain video concept detection

Xu Shaoxi, Yang Jing, Tang Sheng, Zhang Yong-Dong
2011 Proceedings of the Third International Conference on Internet Multimedia Computing and Service - ICIMCS '11  
Due to the mismatch of data distribution between training and testing data set, the issue of semantic gap in the field of video concept detection becomes more and more serious.  ...  Then, these pseudo samples are integrated into the process of Tradboost based cross domain transfer learning to make the best use of semantic information generalized by existing source models.  ...  A lot of cross domain methods have been applied in the field of video concept detection.  ... 
doi:10.1145/2043674.2043681 dblp:conf/icimcs/XuYTZ11 fatcat:aqk657mghfc4jddvq32f7gm7ra

Visual experience recognition using adaptive support vector machine

Santhoshkumar SP, Kumar M Praveen, Beaulah H Lilly
2021 Trends in Computer Science and Information Technology  
between videos from two domains web video domain and consumer video domain.With the help of MATLAB Simulink videos are divided and compared with web domain videos.  ...  Consumer videos are generally captured by amateurs using handheld cameras of events and it contains considerable camera motion, occlusion, cluttered background, and large intraclass variations within the  ...  A cross-domain learning method, Adaptive Multiple Kernel Learning (A-MKL), is used to cope with the considerable variation in feature distributions between videos from the web video domain and consumer  ... 
doi:10.17352/tcsit.000043 fatcat:jpynwp6tbvclhbvwvi6yhgrgtu

Domain Transfer SVM for video concept detection

Lixin Duan, I.W. Tsang, Dong Xu, S.J. Maybank
2009 2009 IEEE Conference on Computer Vision and Pattern Recognition  
To cope with the tremendous change of feature distribution between different domains in video concept detection, we propose a new cross-domain kernel learning method.  ...  Our method, referred to as Domain Transfer SVM (DTSVM), simultaneously learns a kernel function and a robust SVM classifier by minimizing the both structural risk functional of SVM and distribution mismatch  ...  [20] proposed Adaptive SVM (A-SVM) to enhance the prediction performance of video concept detection , in which the new SVM classifier f T (x) is adapted from an existing classifier f A (x) trained from  ... 
doi:10.1109/cvprw.2009.5206747 fatcat:4af3gq6rgnfwdoh3ajjopbyqba

Domain Transfer SVM for video concept detection

Lixin Duan, Ivor W. Tsang, Dong Xu, Stephen J. Maybank
2009 2009 IEEE Conference on Computer Vision and Pattern Recognition  
To cope with the tremendous change of feature distribution between different domains in video concept detection, we propose a new cross-domain kernel learning method.  ...  Our method, referred to as Domain Transfer SVM (DTSVM), simultaneously learns a kernel function and a robust SVM classifier by minimizing the both structural risk functional of SVM and distribution mismatch  ...  [20] proposed Adaptive SVM (A-SVM) to enhance the prediction performance of video concept detection , in which the new SVM classifier f T (x) is adapted from an existing classifier f A (x) trained from  ... 
doi:10.1109/cvpr.2009.5206747 dblp:conf/cvpr/DuanTXM09 fatcat:kavhwe2s6vdillh6pn275dh4hq

A framework for classifier adaptation and its applications in concept detection

Jun Yang, Alexander G. Hauptmann
2008 Proceeding of the 1st ACM international conference on Multimedia information retrieval - MIR '08  
In the experiments of adapting semantic concept detectors across video channels/types, our adaptation approach is proven to be superior to using original (unadapted) classifiers or building new ones in  ...  There is often a need to adapt supervised classifiers such as semantic concept detectors across different domains of data.  ...  EXPERIMENTS IN CROSS-DOMAIN CONCEPT DETECTION Data Collections We use the development set of the TRECVID benchmark video collection in 2005 and 2007, referred to as TV05Dev and TV07Dev.  ... 
doi:10.1145/1460096.1460171 dblp:conf/mir/YangH08 fatcat:jpdxiq473fhbffdu4avicwbahy

(Un)Reliability of video concept detection

Jun Yang, Alexander G. Hauptmann
2008 Proceedings of the 2008 international conference on Content-based image and video retrieval - CIVR '08  
Great effort has been made to improve video concept detection and continuous progress has been reported.  ...  Adopting a more rigorous evaluation approach, we find that most concept classifiers built using the mainstream approach are unreliable because they generalize poorly to domains other than their training  ...  The basic approach of concept detection is to use classification algorithms, typically support vector machines (SVMs), to build concept classifiers which predict the relevance between images or video shots  ... 
doi:10.1145/1386352.1386367 dblp:conf/civr/YangH08 fatcat:llydvlk7lrhjljbdc7fk2tcnly

Exploring multi-modality structure for cross domain adaptation in video concept annotation

Shaoxi Xu, Sheng Tang, Yongdong Zhang, Jintao Li, Yan-Tao Zheng
2012 Neurocomputing  
Domain adaptive video concept detection and annotation has recently received significant attention, but in existing video adaptation processes, all the features are treated as one modality, while multimodalities  ...  To our best knowledge, it is the first time to introduce multi-modality transfer into the field of domain adaptive video concept detection and annotation.  ...  In the field of video concept detection, domain adaptation techniques are also exploited to tackle the domain change among different video corpus to learn robust classifiers for concept detection in the  ... 
doi:10.1016/j.neucom.2011.05.041 fatcat:dog64kb5qvafxfxzgs6f2y3weq

Effective semantic classification of consumer events for automatic content management

Wei Jiang, Alexander C. Loui
2009 Proceedings of the first SIGMM workshop on Social media - WSM '09  
Various ELFs are generated from different types of elementary-level features by using both cross-domain and within-domain learning: crossdomain approaches use two sets of concept scores at both image and  ...  We study semantic event classification in the consumer domain by incorporating cross-domain and within-domain learning.  ...  The data adaptation approach, e.g., Cross-Domain SVM [12] , selectively uses the most important data from the outside domain to help build the new classifier.  ... 
doi:10.1145/1631144.1631153 dblp:conf/mm/0001L09 fatcat:m62pp7ndhvcclivxqlqw2jik5u

TRECVID 2011 Semantic Indexing Task By NTT-SL-ZJU

Yongqing Sun, Go Irie, Takashi Satou, Akira Kojima, Kyoko Sudo, Masashi Morimoto, Akisato Kimura, Zhihua Zhang
2011 TREC Video Retrieval Evaluation  
In order to utilize the semantic properties of TRECVid target data to make the selected web data more adaptive to the target domain, we propose introducing pseudo relevance feedback (PRF) into the automatic  ...  Invariant Feature Transform (SIFT) descriptors, we focus more on selecting web data for training set construction due to the challenges imposed by the serious distribution of data mismatching in different domains  ...  This is similar to the method proposed in [24] for use in within-domain TRECVid concept detection.  ... 
dblp:conf/trecvid/SunISKSMKZ11 fatcat:hqggu2lrm5fgdfwclqxkywnjwq

VIREO @ TRECVID 2012: Searching with Topology, Recounting will Small Concepts, Learning with Free Examples

Wei Zhang, Chun Chet Tan, Shiai Zhu, Ting Yao, Lei Pang, Chong-Wah Ngo
2012 TREC Video Retrieval Evaluation  
Compared with vireo dtcv, we do not use video level fusion for this run. -F X NO vireo dto 1: Spatial matching with DT by using only the ROI region containing the object.  ...  The vireo group participated in four tasks: instance search, multimedia event recounting, multimedia event detection, and semantic indexing.  ...  Among the 120 concepts performed in domain adaptation runs, only one concept "basketball" is evaluated by NIST. Thus the overall results of three runs using A-SVM are similar to our baseline.  ... 
dblp:conf/trecvid/0031TZYPN12 fatcat:5vplqjsezvhmhl7tcu3h2pwipm

Predicting domain adaptivity

Ting Yao, Chong-Wah Ngo, Shiai Zhu
2012 Proceedings of the 20th ACM international conference on Multimedia - MM '12  
Experimental results show that the prediction accuracy of over 75% can be achieved when transferring concept classifiers learnt from LSCOM (news video domain), ImageNet (Web image domain) and Flickr-SF  ...  (weakly tagged Web image domain), respectively, to TRECVID 2011 dataset (Web video domain).  ...  In the literature, there have been several techniques, in-cluding Adaptive SVMs (A-SVMs) [10] and Domain Transfer SVM (DTSVM) [3] , being proposed for addressing the challenge of cross-domain learning  ... 
doi:10.1145/2393347.2396321 dblp:conf/mm/YaoNZ12 fatcat:wbhzwwbh35czjmmvjtedgew5de
« Previous Showing results 1 — 15 out of 6,982 results