The Experts below are selected from a list of 312 Experts worldwide ranked by ideXlab platform
John H. L. Hansen - One of the best experts on this subject based on the ideXlab platform.
-
time Frequency Correlation based missing feature reconstruction for robust speech recognition in band restricted conditions
IEEE Transactions on Audio Speech and Language Processing, 2009Co-Authors: John H. L. HansenAbstract:Band-limited speech represents one of the most challenging factors for robust speech recognition. This is especially true in supporting audio corpora from sources that have a range of conditions in spoken document retrieval requiring effective automatic speech recognition. The missing-feature reconstruction method has a problem when applied to band-limited speech reconstruction, since it assumes the observations in the unreliable regions are always greater than the latent original clean speech. The approach developed here depends only on reliable components to calculate the posterior probability to mitigate the problem. This study proposes an advanced method to effectively utilize the Correlation information of the spectral components across time and Frequency axes in an effort to increase the performance of missing-feature reconstruction in band-limited conditions. We employ an F1 area window and cutoff border window in order to include more knowledge on reliable components which are highly correlated with the cutoff Frequency band. To detect the cutoff regions for missing-feature reconstruction, blind mask estimation is also presented, which employs the synthesized band-limited speech model without secondary training data. Experiments to evaluate the performance of the proposed methods are accomplished using the SPHINX3 speech recognition engine and the TIMIT corpus. Experimental results demonstrate that the proposed time-Frequency (TF) Correlation based missing-feature reconstruction method is significantly more effective in improving band-limited speech recognition accuracy. By employing the proposed TF-missing feature reconstruction method, we obtain up to 14.61% of average relative improvement in word error rate (WER) for four available bandwidths with cutoff frequencies 1.0, 1.5, 2.0, and 2.5 kHz, respectively, compared to earlier formulated methods. Experimental results on the National Gallery of the Spoken Word (NGSW) corpus also show the proposed method is effective in improving band-limited speech recognition in real-life spoken document conditions.
-
Time-Frequency Correlation based missing-feature reconstruction for robust speech recognition in background noise conditions
2009 Conference Record of the Forty-Third Asilomar Conference on Signals Systems and Computers, 2009Co-Authors: Wooil Kim, John H. L. HansenAbstract:This study proposes a novel missing-feature reconstruction method to improve speech recognition in background noise environments. In order to improve the existing missing-feature reconstruction method which utilizes only Frequency Correlation, a temporal spectral feature analysis is employed to leverage temporal Correlation across neighboring frames. The final estimates for missing-feature reconstruction are obtained by a selective combination of the Frequency Correlation based method and the proposed temporal Correlation based method. Performance of the proposed method is evaluated using the Aurora 2.0 framework with car noise and speech babble conditions. Experimental results demonstrate that the proposed method is more effective at increasing speech recognition performance in adverse conditions. By employing the proposed temporal-Frequency based reconstruction method with SNR-based mask estimation, +21.31% and +20.73% average relative improvements in WER are obtained for car and speech babble conditions, compared to the original Frequency Correlation based method.
Fumiaki Maehara - One of the best experts on this subject based on the ideXlab platform.
-
VTC Spring - Control Overhead Reduction Method Employing Frequency Correlation for MU-MIMO-OFDM THP with User Scheduling
2018 IEEE 87th Vehicular Technology Conference (VTC Spring), 2018Co-Authors: Yukiko Shimbo, Nobuhiro Hiruma, Hirofumi Suganuma, Fumiaki MaeharaAbstract:This paper proposes a method for control overhead reduction by employing Frequency Correlation for MU-MIMO-OFDM THP with semi-orthogonal user selection (SUS). The feature of the proposed approach is to reduce the control overhead generated by subcarrier-based user selection results by limiting the sub-carriers according to the Frequency Correlation. The motivation behind this approach comes from the fact that the same user selection results could be applicable to neighboring subcarriers by increasing the Frequency Correlation. A simple theoretical formula containing a parameter for a power delay profile is adopted to facilitate the derivation of the Frequency Correlation. Frequency Correlation is achieved by generating the power delay profile by averaging the channel state information (CSI) over multiple transmit antennas with the aid of MU-MIMO. Moreover, our efficient method, which ensures the benefit of ordering by reusing the SUS results for the ordering calculation, is adopted to coordinate SUS and THP ordering. The effectiveness of the proposed method is demonstrated in terms of both system capacity and control overhead reduction by comparing it with the subcarrier-based approach.
-
VTC Spring - Efficient Ordering Calculation Method Using Frequency Correlation for MU-MIMO-OFDM THP
2017 IEEE 85th Vehicular Technology Conference (VTC Spring), 2017Co-Authors: Nobuhiro Hiruma, Tomoki Maruko, Hiromichi Tomeba, Takashi Onodera, Fumiaki MaeharaAbstract:This paper proposes an efficient ordering calculation method employing Frequency Correlation in multiuser multiple-input multiple-output orthogonal Frequency division multiplexing (MU-MIMO-OFDM) Tomlinson Harashima Precoding (THP). Although the ordering process is effective to improve the transmission performance of THP, it requires a large amount of computational effort due to sub-carrier-based operation. To cope with this problem, the proposed method thins out the sub-carriers conducting the ordering process based on the Frequency Correlation. The basic idea behind this proposal is that the ordering results are considered to be the same among the neighboring sub-carriers in a relatively small delay spread. Moreover, the Frequency Correlation is derived from a delay profile which can be determined by multiple transmit antennas. The effectiveness of the proposed method is demonstrated in terms of calculation cost and the bit error rate (BER) by comparing with the traditional MU-MIMO-OFDM THP without thinning.
M D Fayer - One of the best experts on this subject based on the ideXlab platform.
-
cls next gen accurate Frequency Frequency Correlation functions from center line slope analysis of 2d Correlation spectra using artificial neural networks
Journal of Physical Chemistry A, 2020Co-Authors: David J Hoffman, M D FayerAbstract:The center line slope (CLS) observable has become a popular method for characterizing spectral diffusion dynamics in two-dimensional (2D) Correlation spectroscopy because of its ease of implementation, robustness, and clear theoretical relationship to the Frequency-Frequency Correlation function (FFCF). The FFCF relates the Frequency fluctuations of an ensemble of chromophores to coupled bath modes of the chemical system and is used for comparison to molecular dynamics simulations and for calculating 2D spectra. While in the appropriate limits, the CLS can be shown to be the normalized FFCF, from which the full FFCF can be obtained, in practice the assumptions that relate the CLS to the normalized FFCF are frequently violated. These violations are due to the presence of homogeneous broadening and motional narrowing. The generalized problem of relating the CLS to the FFCF is reanalyzed by introducing a new set of dimensionless parameters for both the CLS and FFCF. A large data set was generated of CLS parameters derived from numerically modeled 2D line shapes with known FFCF parameters. This data set was used to train feedforward artificial neural networks that act as functions, which take the CLS parameters as inputs and return FFCF parameters. These neural networks were deployed in an algorithm that is able to quickly and accurately determine FFCF parameters from experimental CLS parameters and the fwhm of the absorption line shape. The method and necessary inputs to accurately obtain the FFCF from the CLS are presented.
-
separation of experimental 2d ir Frequency Frequency Correlation functions into structural and reorientation induced contributions
Journal of Chemical Physics, 2015Co-Authors: Patrick L Kramer, Jun Nishida, M D FayerAbstract:A vibrational transition Frequency can couple to its environment through a directional vector interaction. In such cases, reorientation of the vibrational transition dipole (molecular orientational relaxation) and its Frequency fluctuations can be strongly coupled. It was recently shown [Kramer et al., J. Chem. Phys. 142, 184505 (2015)] that differing Frequency-Frequency Correlation function (FFCF) decays, due to reorientation-induced spectral diffusion (RISD), are observed with different two-dimensional infrared polarization configurations when such strong coupling is present. The FFC functional forms were derived for the situation in which all spectral diffusion is due to reorientational motion. We extend the previous theory to include vibrational Frequency evolution (spectral diffusion) caused by structural fluctuations of the medium. Model systems with diffusive reorientation and several regimes of structural spectral diffusion rates are analyzed for first order Stark effect interactions. Additionally, the transition dipole reorientational motion in complex environments is frequently not completely diffusive. Several periods of restricted angular motion (wobbling-in-a-cone) may precede the final diffusive orientational randomization. The polarization-weighted FFCF decays are presented in this case of restricted transition dipole wobbling. With these extensions to the polarization-dependent FFCF expressions, the structural spectral diffusion dynamics of methanol in the room temperature ionic liquid 1-hexyl-3-methylimidazolium hexafluorophosphate can be separated quantitatively from RISD using the experimental center line slope data. In addition, prior results on the spectral diffusion of water, methanol, and ethanol in 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide are re-examined to elucidate the influence of reorientation on the data, which were interpreted in terms of structural fluctuations.
-
Frequency Frequency Correlation functions and apodization in two dimensional infrared vibrational echo spectroscopy a new approach
Journal of Chemical Physics, 2007Co-Authors: Kyungwon Kwak, Sungnam Park, Ilya J Finkelstein, M D FayerAbstract:Ultrafast two-dimensional infrared (2D-IR) vibrational echo spectroscopy can probe structural dynamics under thermal equilibrium conditions on time scales ranging from femtoseconds to ∼100ps and longer. One of the important uses of 2D-IR spectroscopy is to monitor the dynamical evolution of a molecular system by reporting the time dependent Frequency fluctuations of an ensemble of vibrational probes. The vibrational Frequency-Frequency Correlation function (FFCF) is the connection between the experimental observables and the microscopic molecular dynamics and is thus the central object of interest in studying dynamics with 2D-IR vibrational echo spectroscopy. A new observable is presented that greatly simplifies the extraction of the FFCF from experimental data. The observable is the inverse of the center line slope (CLS) of the 2D spectrum. The CLS is the inverse of the slope of the line that connects the maxima of the peaks of a series of cuts through the 2D spectrum that are parallel to the Frequency a...
David L. Bradley - One of the best experts on this subject based on the ideXlab platform.
-
on the relationship between signal bandwidth and Frequency Correlation for ocean surface forward scattered signals
Journal of the Acoustical Society of America, 2005Co-Authors: Lee R Culver, David L. BradleyAbstract:The relationship between the bandwidth of a signal and the Correlation of that signal with its ocean surface reflected arrival, a quantity we term Frequency Correlation, has been investigated experimentally and compared with two theories. DeCorrelation of wideband surface scattered signals is a direct consequence of time spread. The acoustic measurement utilized a very short pure tone signal, from which time spread has been estimated, and four broadband signals with different bandwidths, for which Correlation with the transmitted signal has been measured. An environment-driven model developed by Dahl was used to predict time spread, which agreed favorably with our time spread measurements. The model was also employed in two theories that predict Frequency Correlation. The first, a theory published by Reeves in 1974, is based upon the ratio of signal temporal resolution to total time spread. This theory compared well with our measurements for 1 kHz bandwidth signals, but is not applicable for signal bandwi...
-
On the relationship between signal bandwidth and Frequency Correlation for ocean surface forward scattered signals
The Journal of the Acoustical Society of America, 2005Co-Authors: R. Lee Culver, David L. BradleyAbstract:The relationship between the bandwidth of a signal and the Correlation of that signal with its ocean surface reflected arrival, a quantity we term Frequency Correlation, has been investigated experimentally and compared with two theories. DeCorrelation of wideband surface scattered signals is a direct consequence of time spread. The acoustic measurement utilized a very short pure tone signal, from which time spread has been estimated, and four broadband signals with different bandwidths, for which Correlation with the transmitted signal has been measured. An environment-driven model developed by Dahl was used to predict time spread, which agreed favorably with our time spread measurements. The model was also employed in two theories that predict Frequency Correlation. The first, a theory published by Reeves in 1974, is based upon the ratio of signal temporal resolution to total time spread. This theory compared well with our measurements for 1 kHz bandwidth signals, but is not applicable for signal bandwidths greater than about 2 kHz. The second, a theory developed by Ziomek, models ocean acoustic propagation as transmission through a linear system. This theory agreed well with our Frequency Correlation measurements for signal bandwidths of 1–22 kHz.
-
On the Relationship between Signal Bandwidth and Frequency Correlation for Surface Forward Scattered Signals
AIP Conference Proceedings, 2004Co-Authors: Lee Culver, David L. BradleyAbstract:The relationship between the signal bandwidth and the Correlation of a single surface reflected arrival with the transmitted signal has been investigated experimentally and compared with two theories. The dependence of Correlation on signal bandwidth is termed Frequency Correlation. DeCorrelation of surface scattered signals is a direct consequence of time spread. Thus the acoustic measurement utilized two pure tone signals, from which time spread has been estimated, and four broadband signals with different bandwidths, from which Correlation with the transmitted signal has been calculated. A model developed by Dahl for the ocean surface bistatic scattering cross section was used to predict time spread, which agreed very well with the measured time spread. Next, scattering cross section prediction was employed in two theories that predict Frequency Correlation. The first, published by Reeves in 1974, compared well with the measurements for bandwidths up to 2 kHz, but under predicted Correlation for signal bandwidth between 7 and 22 kHz. In the second, linear systems theory was used to develop a mathematical relationship between time spread and Frequency Correlation. Predictions made using the linear systems theory agree well with the measured values for signal bandwidths up to 22kHz. Further work is required to evaluate the linear systems theory under higher sea state conditions.
Chandra Athaudage - One of the best experts on this subject based on the ideXlab platform.
-
A framework to calculate space-Frequency Correlation in multi-carrier systems
IEEE Transactions on Wireless Communications, 2010Co-Authors: Tharaka A. Lamahewa, Thushara D. Abhayapala, Rauf Iqbal, Chandra AthaudageAbstract:This paper proposes a general framework for the calculation of space-Frequency Correlation at the receiver in time-invariant Frequency selective channel environments. The proposed framework incorporates antenna locations to capture the spatial information of the channel and a general joint angle-delay power distribution to capture the Frequency selectivity of the multipath channel. To demonstrate the strength of the proposed framework, we apply it on an orthogonal Frequency division multiplexing (OFDM) system to obtain the space-Frequency Correlation between two antenna elements for two analytically derived joint angle-delay power distributions.