Search
Search Funnelback University
- Refined by:
- Date: 2006
1 -
50 of
59
search results for news |u:mi.eng.cam.ac.uk
Fully-matching results
-
Royal Society Meeting on Geometry in Computer Vision
mi.eng.cam.ac.uk/~cipolla/royal_society.html28 Nov 2006: Discussion Meeting on. New Geometric Techniques in Computer Vision. Wednesday 23 July and Thursday 24 July 1997. ... The adoption of projective geometry in the mid-1980's, as a supplement to Euclidean geometry, provided a new approach to many computer -
Who Really Spoke When? Finding Speaker Turns and Identities in…
mi.eng.cam.ac.uk/reports/full_html/tranter_icassp06.html/9 Dec 2006: Who Really Spoke When? Finding Speaker Turns and Identities in Broadcast News Audio. ... For example, {[name] <SHOWNAME> <PLACE>} may be a more powerful predictor than just {[name] BBC News Washington}. -
THE CU-HTK MANDARIN BROADCAST NEWS TRANSCRIPTION SYSTEM R. Sinha, ...
mi.eng.cam.ac.uk/~mjfg/sinha_ICASSP06.pdf22 Nov 2006: ABSTRACT. This paper discusses the development of the CU-HTK MandarinBroadcast News (BN) transcription system. ... Jin, M. Noamany, andT. Shultz, “The ISL RT-04 Mandarin Broadcast News evaluationsystem,” inProc. -
THE CU-HTK MANDARIN BROADCAST NEWS TRANSCRIPTION SYSTEM R. Sinha, ...
mi.eng.cam.ac.uk/research/projects/AGILE/publications/rs_ICASSP06.pdf23 Feb 2006: Jin, M. Noamany, andT. Shultz, “The ISL RT-04 Mandarin Broadcast News evaluationsystem,” inProc. ... Xu, “The BBN MandarinBroadcast News transcription system,” inProc. Eurospeech,Lisbon, Portugal, September 2005. -
Machine Learning for Speech & LanguageProcessing Mark Gales 28 ...
mi.eng.cam.ac.uk/~mjfg/FCSW_talk.pdf19 Jul 2006: Text data: used to train the ASR language model:– large news corpora available;– systems built on > 1 billion words of data. • -
A New Look at Filtering Techniques for Illumination Invariance ...
mi.eng.cam.ac.uk/reports/svr-ftp/arandjelovic_AFG06.pdf30 Jan 2006: A New Look at Filtering Techniques for Illumination Invariance in AutomaticFace Recognition. -
Joint Uncertainty Decoding for Noise Robust Speech Recognition H. ...
mi.eng.cam.ac.uk/~mjfg/liao_INTER05.pdf19 Dec 2006: certainty decoding framework is presented, Joint uncertaintydecoding. This new approach is compared to both the stan-dard and uncertainty versions of SPLICE. ... Frey, “Accounting for uncertaintyin observations: A new paradigm for robust speech -
Explicitly Generating Complementary Systems for Large…
mi.eng.cam.ac.uk/~mjfg/breslin_INTER06.pdf22 Nov 2006: This paper presents a new algorithm, Minimum BayesRisk Leveraging (MBRL), for explicitly generating systems thatare complementary to each other. ... and Tranter, S.E., “Progress in the CU-HTKBroadcast News Transcription System”, IEEE Trans. -
DEVELOPMENT OF THE CUHTK 2004 MANDARIN CONVERSATIONAL TELEPHONESPEECH …
mi.eng.cam.ac.uk/~mjfg/gales_ICASSP05.pdf22 Nov 2006: The total contributionfrom all the news corpora was about 0.12, with the majority fromPeople’s Daily (0.09). ... Monkowski, M. A. Picheny,and K. Shen, “New methods in continuous Mandarin speechrecognition,” inProc. -
C:/SFWDoc/Academic/Publications/2006/ICPR_2006/Final_ContGest/icpr_200…
mi.eng.cam.ac.uk/reports/svr-ftp/sfwong_icpr06a.pdf21 Sep 2006: 3. Feature extraction is done within a. t1. on the new weights (i.e. ... 5 Conclusion. A new method is proposed to recognise continuouselementary gestures in this paper and it can be applied tosign language recognition. -
C:/SFWDoc/Academic/Publications/2006/ICPR_2006/Final_AppTrack/icpr_200…
mi.eng.cam.ac.uk/reports/svr-ftp/sfwong_icpr06b.pdf21 Sep 2006: This meanswe can start from an initially small model and test the ‘rele-vance’ of each new input vectori sequentially. ... 5 Conclusion. In this paper, a new appearance-based method is pro-posed to track an object which may undergo appearancechange. -
University of CambridgeEngineering Part IB Information Engineering…
mi.eng.cam.ac.uk/~cipolla/lectures/PartIB/old/2006-IB-search.pdf19 May 2006: return new Node(value, nothing, nothing);. if(value == current.value). return current;. if(value < current.value). ... list<vector> leftData = new list<vector>();. list<vector> rightData = new list<vector>();. foreach(vector dataPoint in data). {. -
TechReport.dvi
mi.eng.cam.ac.uk/reports/svr-ftp/ng_tr529.pdf21 Apr 2006: To demonstrate this formally, we introduce in Figure 1(b) a new setof axes labelled xp and yp which are simply the x and y axes rotated anticlockwise through an3. -
paper.dvi
mi.eng.cam.ac.uk/~mjfg/liao_INTER06.pdf22 Nov 2006: in observations: A new paradigm for robust speech recogni-tion,” in Proc. -
AUGMENTED STATISTICAL MODELS FOR SPEECH RECOGNITION M.I. Layton and…
mi.eng.cam.ac.uk/~mjfg/layton_ICASSP06.pdf22 Nov 2006: ABSTRACTRecently there has been significant interest in developing new acous-tic models for speech recognition. ... statistics are introduced (only new functions of thebase model statistics), independence assumptions of the base modelare retained. -
EUROGRAPHICS 2006 / E. Gröller and L. Szirmay-Kalos(Guest Editors) ...
mi.eng.cam.ac.uk/reports/svr-ftp/hernandez_eg06.pdf19 Sep 2006: Bottom: proposed method. The result-ing surface is filtered from noise while new high frequencygeometry is revealed (note the reconstructed surface cracks). -
WHO REALLY SPOKE WHEN?FINDING SPEAKER TURNS AND IDENTITIES IN ...
mi.eng.cam.ac.uk/reports/svr-ftp/tranter_icassp06.pdf9 Dec 2006: WHO REALLY SPOKE WHEN?FINDING SPEAKER TURNS AND IDENTITIES IN BROADCAST NEWS AUDIO. ... 1014. 4. EXPERIMENTAL SET UP. The training data used for this task consisted of the Hub-4 1996/7broadcast news training data. -
Incremental Learning of Locally OrthogonalSubspaces for Set-based…
mi.eng.cam.ac.uk/reports/svr-ftp/kim_bmvc06.pdf21 Sep 2006: i). (6). Note this is applied only to the classes having the new data. ... The latter term is for computing (Mni , P. ni , Λni ) from the new data. -
Incremental Learning of Temporally-CoherentGaussian Mixture Models…
mi.eng.cam.ac.uk/reports/svr-ftp/arandjelovic_SME06.pdf14 Mar 2006: Common to both [8] and [14] is the failure to make use of the existing modelwhen the GMM corresponding to new data is fitted. ... c) New data point (red circle) causes the splitting of a Gaussian component,resulting in a 3-component mixture. -
Complementary System Combination andGeneration for ASR Mark Gales 20…
mi.eng.cam.ac.uk/~mjfg/mjfg_tcstar.pdf5 Jul 2006: Sim, and P. C. Woodland, “The CU-HTK Mandarin Broadcast News transcriptionsystem,” in Proceedings ICASSP, 2006. ... Venkataraman, R. R. Gadde, and A. Stolcke, “SuperEARS: Multi-site broadcast news system,” inProc. -
Discriminative Adaptation for Speaker Verification C. Longworth and…
mi.eng.cam.ac.uk/~mjfg/longworth_INTER06.pdf22 Nov 2006: For MMI-MAP the the new estimate of the mean ofcomponent j for speaker s, µ̂(s)j given the current estimate µ. -
C:/SFWDoc/Academic/Publications/2005/ICCV_HCI_2005/FinalPaper/iccv_hci…
mi.eng.cam.ac.uk/reports/svr-ftp/sfwong_iccv_hci05.pdf21 Sep 2006: All RVM classifierswill be trained separately by evaluating the relevance of this new sample. ... online re-training scheme can increase the adaptability of the classifierto input captured under a new condition. -
C:/SFWDoc/Academic/Publications/2005/BMVC_2005/FinalPaper/bmvc_05_sfwo…
mi.eng.cam.ac.uk/reports/svr-ftp/sfwong_bmvc05.pdf21 Sep 2006: Thus, thefirst 17 eigenvectors arechosen as the new basis functions for converting any new incoming MGO image (O) intoa new feature vector. ... 5 Conclusion. A new method is proposed torecognise primitive movementsin this paper and it can beapplied -
TOWARD AUTOMATIC BLOOD SPATTER ANALYSIS INCRIME SCENES A.R. Shen, ...
mi.eng.cam.ac.uk/reports/svr-ftp/brostow_BloodSpatter.pdf14 Sep 2006: There appears to be an angle-dependent bias. Interestingly, Bevel and Gardner [1] point outthat it is “Difficult for a new analyst to differentiate bloodstain. ... In International Association of BloodpatternAnalysis News, pages 4–12, 1990. [10] -
IEEE TRANS. ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2007 ...
mi.eng.cam.ac.uk/~mjfg/gales_ASL.pdf22 Nov 2006: The designof a real-time broadcast news recognition system is then detailedusing these new models. ... Gales et al.: PROGRESS IN THE CU-HTK BROADCAST NEWS TRANSCRIPTION SYSTEM 3. -
Hole Filling Through Photomontage Marta Wilczkowiak∗, Gabriel J.…
mi.eng.cam.ac.uk/reports/svr-ftp/brostow_HoleFillingBMVC05.pdf14 Sep 2006: Thisextends the texture synthesis algorithm of [9] where texture is synthesized patch by patch,each new patch being chosen from the source region based on similarity to the targetneighborhood. ... Consider placing a new patch m, slightly overlapping a -
EUROGRAPHICS 2006 / E. Gröller and L. Szirmay-Kalos(Guest Editors) ...
mi.eng.cam.ac.uk/reports/svr-ftp/johnson_semantic06.pdf1 Jun 2006: We propose a technique that allows a person to design a new photographwith substantially less effort. ... Our primary contribution is a new method forcreating images of specific things and people, with minimalhuman effort. -
WHO REALLY SPOKE WHEN?FINDING SPEAKER TURNS AND IDENTITIES IN ...
mi.eng.cam.ac.uk/reports/full_html/tranter_icassp06.html/paper.pdf9 Dec 2006: WHO REALLY SPOKE WHEN?FINDING SPEAKER TURNS AND IDENTITIES IN BROADCAST NEWS AUDIO. ... 1014. 4. EXPERIMENTAL SET UP. The training data used for this task consisted of the Hub-4 1996/7broadcast news training data. -
Gaze Manipulation for One-to-one Teleconferencing A. Criminisi, J.…
mi.eng.cam.ac.uk/reports/svr-ftp/shotton_iccv03.pdf15 Feb 2006: The second is representingand using the computed disparities and occlusions to gen-erate a new view. ... The new model is illustrated in the diagramin fig. 6 (to be compared with fig. -
JOURNAL OF IEEE TRANS. ACOUST., SPEECH, SIGNAL PROCESSING, JULY ...
mi.eng.cam.ac.uk/~mjfg/sim_SAP06.pdf22 Nov 2006: and K subsumes terms independent of the model parameters.θ denotes the set of new parameter. ... Discriminative training of precision matrices was evaluatedon an English conversational telephone speech (CTS) task,which consists of multi-speaker -
Learning Discriminative Canonical Correlationsfor Object Recognition…
mi.eng.cam.ac.uk/reports/svr-ftp/kim_eccv06.pdf21 Sep 2006: The new technique facilitates effective dis-criminative learning over sets, thus providing an impressive set classification ac-curacy. -
Hierarchical Part-Based Human Body PoseEstimation R. Navaratnam∗ A.…
mi.eng.cam.ac.uk/reports/svr-ftp/navaratnam_hierarchical.pdf14 Sep 2006: We propose a new hierarchical part-based pose estimationmethod for the upper-body that efficiently searches the high dimensional ar-ticulation space. ... C. Wolf. Parametric correspondence and chamfer. matching: Two new techniques for image matching.Proc. -
Sparse and Semi-supervised Visual Mapping with the S3GP Oliver ...
mi.eng.cam.ac.uk/reports/svr-ftp/williams_cvpr06.pdf3 Apr 2006: In [14], error is computed using a“leave-one-out” test rather than with completely new testdata. ... 7] A. Jain. Fundamentals of Digital Image Processing. SystemSciences. Prentice-Hall, New Jersey, 1989. -
techreport_20060422MJ.dvi
mi.eng.cam.ac.uk/reports/svr-ftp/brostow_Eurographics06.pdf14 Sep 2006: Our primary contributionis a new method for creating images of specific thingsand people, with minimal human effort. ... In addition, their in-terface is meant for detailed manipulation of a manuallyspecified region, while our method allows for very -
Automatic Cast Listing in Feature-Length Films with Anisotropic…
mi.eng.cam.ac.uk/reports/svr-ftp/arandjelovic_CVPR06.pdf21 Mar 2006: 6] consider the problem of clustering detectedfrontal faces extracted from web news pages. ... Hart, and D. G. Stork.Pattern Classification. JohnWily & Sons, Inc., New York, 2nd edition, 2000.5. -
pami04.dvi
mi.eng.cam.ac.uk/reports/svr-ftp/hernandez_pami06.pdf19 Sep 2006: 1a). We present a new approach to silhouette-based camera motionand focal length estimation that. ... 6). Increasingδ makes the new measure more robust against bad segmentation.But robustness is. -
Product of Gaussians for Speech Recognition M.J.F. Gales and ...
mi.eng.cam.ac.uk/~mjfg/airey_CSL06.pdf22 Nov 2006: Two new formsof representation are discussed. The first uses MoG as the experts. ... 2] H Bourlard and S Dupont. A new ASR approach based on independentprocessing and combination of partial frequency bands. -
Multi-Sensory Face Biometric Fusion (for Personal Identification)…
mi.eng.cam.ac.uk/reports/svr-ftp/arandjelovic_OTCBVS06.pdf19 Mar 2006: Names and faces inthe news.CVPR, 2004. [6] Å. Björck and G. ... 11] T-J. Chin and D. Suter. A new distance criterion for face recognitionusing image sets.ICCV, pages 549–558, 2006. -
The Layout Consistent Random Field for Recognizing and Segmenting ...
mi.eng.cam.ac.uk/reports/svr-ftp/shotton_cvpr06.pdf3 Apr 2006: It is possible to pre-select which parts are used asin [4], though this requires considerable human effort foreach new object class. ... The unary classifiers arelearned as described above, after which a new labelling isinferred for all the training -
TextonBoost: Joint Appearance, Shape andContext Modeling for…
mi.eng.cam.ac.uk/reports/svr-ftp/shotton_eccv06.pdf15 Feb 2006: wπ. (8). Given this new parameter setting, a new class labeling is inferred and this pro-cedure is iterated [15]. ... 10b with 10d) since the object contour is accurately delineated. 6 ConclusionsThis paper has presented a new discriminative model for -
article.dvi
mi.eng.cam.ac.uk/~mjfg/rosti_CSL04.pdf22 Nov 2006: The cross terms including the new state space mean vectors and thefirst-order accumulates have been simplified in Equation (14). ... The lth row vector ĉjl of the new observation matrix can bewritten as. -
IEICE TRANS. ??, VOL.Exx–??, NO.xx XXXX 200x1 INVITED PAPER ...
mi.eng.cam.ac.uk/~mjfg/gales_IEICE06.pdf21 Nov 2006: M.J.F. GALES† and M.I. LAYTON†, Nonmembers. SUMMARY There has been significant interest in develop-ing new forms of acoustic model, in particular models which al-low additional dependencies -
SUB-SAMPLE INTERPOLATIONSTRATEGIES FOR SENSORLESSFREEHAND 3D…
mi.eng.cam.ac.uk/reports/svr-ftp/housden_tr545.pdf13 Jan 2006: The new technique is shown to be superior, limitinginterpolation errors to around 5% of the length of the freehand reconstruction. -
Semi-supervised Learning of Joint DensityModels for Human Pose…
mi.eng.cam.ac.uk/reports/svr-ftp/navaratnam_semi_supervised.pdf14 Sep 2006: W. Anderson. An Introduction to Multivariate Statistical Analysis. John Wiley & Sons,Inc., New York,3rd edition, 2003. -
SENSORLESS RECONSTRUCTIONOF UNCONSTRAINED FREEHAND 3D ULTRASOUND DATA …
mi.eng.cam.ac.uk/reports/svr-ftp/housden_tr553.pdf22 May 2006: In Section 3, we describe experiments with a hand-held probe designedto test the capabilities and limitations of the new algorithms. ... 5. r (b )half 1. 11 maxb r (b ). ideal new frame, f n. -
Unsupervised Bayesian Detection of Independent Motion in Crowds…
mi.eng.cam.ac.uk/reports/svr-ftp/brostow_MotionInCrowdsCVPR06.pdf14 Sep 2006: These would in turn preventthe “spawning” of new entity hypotheses in the middle ofthe scene. ... References[1] G. Bradski. Opencv: Examples of use and new applications in stereo, recogni-. -
Face Recognition from Video using the GenericShape-Illumination…
mi.eng.cam.ac.uk/reports/svr-ftp/arandjelovic_ECCV06.pdf17 Feb 2006: Compared to single-shot recognition, face recognition from image sequencesis a relatively new area of research. ... PAMI, 19(7):721–732, 1997. 2. O. Arandjelović and R. Cipolla. A new look at filtering techniques for illuminationinvariance in -
Journal of VLSI Signal Processing manuscript No.(will be inserted ...
mi.eng.cam.ac.uk/~mjfg/layton_JVLSI07.pdf22 Nov 2006: These paths represent the transformationof an input sequence into a new output sequence. ... Theonly change required is the addition of a new transducer,S, to calculate the state posteriors. -
stenger_imavis06.dvi
mi.eng.cam.ac.uk/reports/svr-ftp/stenger_imavis06.pdf21 Sep 2006: C.Wolf. Parametric correspondence and chamfer matching:Two new techniques for image matching. ... InARPA Image Un-. derstanding Workshop, pages 1179–1183, New Orleans,LA, May 1997. -
IEEE TRANS. ON SAP, VOL. ?, NO. ??, ????? ...
mi.eng.cam.ac.uk/research/projects/AGILE/publications/mjfg_ASL.pdf23 Feb 2006: The designof a real-time broadcast news recognition system is then detailedusing these new models. ... Gales et al.: THE CUED BROADCAST NEWS TRANSCRIPTION SYSTEM 3. Segmentation /#Segments #Clusters.
Search history
Recently clicked results
Recently clicked results
Your click history is empty.
Recent searches
Recent searches
Your search history is empty.