Search

22 Jul 2010: Achievements. Freehand RF 3D ultrasound acquisition. We have successfully developed the real-time, RF 3D ultrasound acquisition system [24]. ... 24] G. Treece, R. Prager and A. Gee. Freely available software for 3D RF ultrasound.

5 Jan 2010: use hinge-loss [f (x)]. Many variants possible [21, 22, 23, 24]. ... ICML, 2008. [24] G Saon and D Povey, “Penalty function maximization for large margin HMM training,” in Proc.

10 Aug 2010: The individual image frames arelocated in space by a position sensor attached to the probe [16] or, in the case of intravascularimaging, by a continuous pullback method [24]. ... Communications of theACM, 24(6):381–395, June 1981. [10] T. G. Fisher, T.

25 Oct 2010: This, however, requires setting of a learning rate. Ye et al.[24] have proposed an incremental version of LDA, which caninclude a single new data point in each time step. ... California, Dept. Statistics, Berkeley, CA,2005, Tech. Rep. 688. [24] J. Ye, Q.

19 Nov 2010: on splicing 9 time frames of 24 dimensional Cepstra, including c0.A context dependent state–clustered allophone system was builton the broadcast news training data.

12 Jan 2010: 21 0.78464 1.10 1.05 0.47 0.75 0.10 0.54 0.05 0.55465 0.95 1.18 0.24 0.96 0.02 0.68 ... 70 0.22 0.54 0.19 0.54all 1.24 0.83 0.19 0.62 0.01 0.48 0.01 0.47.

8 Sep 2010: Speech Lang.,vol. 24, no. 4, pp. 648–662, 2010. [8] B. Taskar, “Learning structured prediction models: a large marginapproach,” Ph.D.

26 May 2010: DT-CWT [24, 25] which has been shown to be particularly effective in denoising applications [26].

14 Sep 2010: Using 17 pairs, about 24% of thetotal number of pairs, 92% of the WER improvement usingthe 1-v-1 system over the VTS baseline was achieved.

14 Sep 2010: yt µ. (ωjM)y. ). . (24). where the clean models have been compensated for the test condition Y,λ = {λ(ωl)y , λ(ωj)y }, and there are a total of ... 05 8.78 8.43 6.20 9.53 8.22. 00

3 Sep 2010: This chap-ter examines approaches that handle background-noise and channel distortions bymodifying the parameters of the underlying acoustic models, in this case HiddenMarkov Models (HMMs) [52, 24]. ... Though intuitively well motivated, from (24) it can

25 Oct 2010: This, however, requires setting of a learning rate. Ye et al.[24] have proposed an incremental version of LDA, which caninclude a single new data point in each time step. ... California, Dept. Statistics, Berkeley, CA,2005, Tech. Rep. 688. [24] J. Ye, Q.

25 Oct 2010: ZT SB,3Z = RRT T T ZT SB,3Z = RRT. (24).

25 Oct 2010: ZT SB,3Z = RRT T T ZT SB,3Z = RRT. (24).

25 Feb 2010: An earlyexample of this type of kernel is the generalised linear discriminant sequence (GLDS) ker-nel [24, 142]. ... Recent approaches have examined how to applysupport vector machines to the SV task using dynamic kernels [24, 205].

6 Oct 2010: Lattice Rescoring Methods forStatistical Machine Translation. Graeme Blackwood. Cambridge University Engineering Departmentand. Clare College. Dissertation submitted to the University of Cambridgefor the degree of Doctor of Philosophy.

10 Jun 2010: The power. spectrum is then down-sampled using a bank of triangular filters, with typically 24 to 40.

20 Oct 2010: 242.8.3 Other evaluation criteria. 24. 2.9 Pre-processing of data. 252.9.1 Tokenisation and morphological analysis of Arabic text.