epartement´Electronique et Physique, Institut National des T´el´ecommunications, 9 rue Charles Fourier,91011 Evry, FranceReceived 1 May 2006; Revised 9 October 2006; Accepted 20 November 2006Recommended by William Allan SandhamThis work aims at providing new insights on the electrocardiogram (ECG) s[r]
the residual of the reconstruction of the image using thestate's eigenvectors. The state membership is thus invariantto variance along the eigenvectors. Although not applied toimages directly, the present work is an extension of thisWILSON AND BOBICK: PARAMETRIC HIDDEN MARKOV MODELS FO[r]
dimensional array of intensity values, which is comparedwith other facial arrays. Techniques of this type include prin-cipal component analysis (PCA) [6], where the varianceamong a set of face images is represented by a number ofeigenfaces. The face images, encoded as weight vectors of theeig[r]
long-range dependency problem inherent to traditional HMMs has been drastically reduced. SHMMs have not previously beenapplied to the problem of face identification. The results reported in this application have shown that SHMM outperforms thetraditional hidden Markov model with[r]
cally, we introduce a novel architecture that enables real-timespeech recognition on an FPGA utilizing the 90 nm ASICmultiply-accumulate and block RAM features of the XilinxVirtex 4 series devices. Final conclusions as well as a sum-mary of synthesis and post place-and-route results wi[r]
ON MERGING HIDDEN MARKOV MODELS WITH DEFORMABLE TEMPLATES Ram R. Rao and Russell M. Mersereau School of Electrical and Computer Engineering Georgia Institute of Technology Atlanta, Georgia 30332 rr@eedsp.gatech.edu ABSTRACT Hidden Markov modeling has proven extremely usef[r]
characterized dehydrogenases ⁄ reductases showed aspacing of these glycine residues in a GxGxxG pattern,where ‘x’ denotes any residue [5,6]. However, as newmembers of this fold have been recognized, the generalpattern is now described as Gx(x)Gx(x)G [7], i.e. thespacing between the glycine re[r]
sional output pdf vectorand the initial state distribu-tion vectorwhich consists of the probabilities. After the model has been trained us-ing the Baum-Welch algorithm, feature sequencescan be scored according to(3)Usually the likelihoodis estimated by the Viterbialgorithm, which is an approx[r]
set of features and in order to be able to compareHMM and CRF models under the same c onditionswe have not applied such commonly used featuresas orthografic or morphological ones. The only ad-ditional information we have exploited are parts-of-sp e e ch (POS) tags.The set of POS tags was supplied by[r]
can be computed in a HMM way by adding to each state ofthe model an observation probability of the input .6. ConclusionA new hand gesture recognition method based on In–put/Output Hidden Markov Models is presented. IOHMMdeal with the dynamic aspects of gestures. They have[r]
1. Liu, W, Wang, LY: The Markov approximation of the random field on Cayley tree and a class of small deviationtheorems. Stat Probab Lett. 63, 113–121 (2003). doi:10.1016/S0167-7152(03)00058-02. Shi, ZY, Yang, WG: Some limit properties for the mth-order nonhomogeneous Markov<[r]
recognition since the 1970s and in bioinformatics since the1990s. In automated gene finding, there are two types of ap-proaches based on data intrinsic to the genome under studyor extrinsic to the genome (e.g., homology and EST data).Since around 2000, the best gene finders have been basedon combined[r]
sistent with a model of human entity generation,but this direction of causality is chosen to representthe effect of semantics (referents) generating syn-tax (POS tags). In addition, this is a joint model inwhich POS tagging and coreference resolution areintegrated togethe[r]
c and others, 2006).On the tectogrammatical layer, each sentence isrepresented as a tectogrammatical tree (t-tree forshort; abbreviations t-node and t-layer are used inthe further text too). The main features of t-trees(from the viewpoint of our experiments) are fol-lowing. Each[r]
Hindawi Publishing CorporationEURASIP Journal on Audio, Speech, and Music ProcessingVolume 2009, Article ID 497292, 9 pagesdoi:10.1155/2009/497292Research ArticleDrum Sound Detection in Polyphonic Music withHidden Markov ModelsJouni Paulus and Anssi KlapuriDepartment of Signal Processing, Tam[r]
, Unik – Uni. Graduate center. [3]. Andrew Adam, Saleema Amershi, Identifying humans by their walk and generating new motions using hidden Markov models, CPSC 532A Topics in AI: Graphical models and CPSC 526 computer animation, December 15, 2004. [4]. Lawrence Rabiner and Biing-[r]
detection performance on the TIMIT dataset. In thefuture, we plan to explore phonological context anduse more flexible topological structures to modelacoustic units within our framework.AcknowledgementsThe authors would like to thank Hung-an Chang andEkapol Chuangsuwanich for training the Englishand[r]
MT has received reasonable attention. For exam-ple, Goldwater & McClosky (2005) show improve-ments when preprocessing Czech input to reflecta morphological decomposition using combinationsof lemmatization, pseudowords, and morphemes.Yeniterzi and Oflazer (2010) bridge the morpholog-ical dispar[r]