摘要 :
In recent years, a number of feature extraction procedures for automatic speech recognition (ASR) systems have been based on models of human auditory processing, and one often hears arguments in favor of implementing knowledge of ...
展开
In recent years, a number of feature extraction procedures for automatic speech recognition (ASR) systems have been based on models of human auditory processing, and one often hears arguments in favor of implementing knowledge of human auditory perception and cognition into machines for ASR. This paper takes a reverse route, and argues that the engineering techniques for automatic recognition of speech that are already in widespread use are often consistent with some well-known properties of the human auditory system.
收起
摘要 :
In speaker diarization, standard approaches typically perform speaker clustering on some initial segmentation before refining the segment boundaries in a re-segmentation step to obtain a final diarization hypothesis. In this paper...
展开
In speaker diarization, standard approaches typically perform speaker clustering on some initial segmentation before refining the segment boundaries in a re-segmentation step to obtain a final diarization hypothesis. In this paper, we integrate an improved clustering method with an existing re-segmentation algorithm and, in iterative fashion, optimize both speaker cluster assignments and segmentation boundaries jointly. For clustering, we extend our previous research using factor analysis for speaker modeling. In continuing to take advantage of the effectiveness of factor analysis as a front-end for extracting speaker-specific features (i.e., i-vectors), we develop a probabilistic approach to speaker clustering by applying a Bayesian Gaussian Mixture Model (GMM) to principal component analysis (PCA)-processed i-vectors. We then utilize information at different temporal resolutions to arrive at an iterative optimization scheme that, in alternating between clustering and re-segmentation steps, demonstrates the ability to improve both speaker cluster assignments and segmentation boundaries in an unsupervised manner. Our proposed methods attain results that are comparable to those of a state-of-the-art benchmark set on the multi-speaker CallHome telephone corpus. We further compare our system with a Bayesian nonparametric approach to diarization and attempt to reconcile their differences in both methodology and performance.
收起
摘要 :
A low-noise wideband receiver (Rx) is presented for a multichannel wireless implantable neural recording (WINeR) system that utilizes time-division multiplexing of pulse width modulated (PWM) samples. The WINeR-6 Rx consists of fo...
展开
A low-noise wideband receiver (Rx) is presented for a multichannel wireless implantable neural recording (WINeR) system that utilizes time-division multiplexing of pulse width modulated (PWM) samples. The WINeR-6 Rx consists of four parts: 1) RF front end; 2) signal conditioning; 3) analog output (AO); and 4) field-programmable gate array (FPGA) back end. The RF front end receives RF-modulated neural signals in the 403–490 MHz band with a wide bandwidth of 18 MHz. The frequency-shift keying (FSK) PWM demodulator in the FPGA is a time-to-digital converter with 304 ps resolution, which converts the analog pulse width information to 16-bit digital samples. Automated frequency tracking has been implemented in the Rx to lock onto the free-running voltage-controlled oscillator in the transmitter (Tx). Two antennas and two parallel RF paths are used to increase the wireless coverage area. BCI-2000 graphical user interface has been adopted and modified to acquire, visualize, and record the recovered neural signals in real time. The AO module picks three demultiplexed channels and converts them into analog signals for direct observation on an oscilloscope. One of these signals is further amplified to generate an audio output, offering users the ability to listen to ongoing neural activity. Bench-top testing of the Rx performance with a 32-channel WINeR-6 Tx showed that the input referred noise of the entire system at a Tx–Rx distance of 1.5 m was 4.58 μV$_{rm rms}$ with 8-bit resolution at 640 kSps. In an in vivo experiment, location-specific receptive fields of hippocampal place cells were mapped during a behavioral experiment in which a rat completed 40 laps in a large circular track. Results were compared against those acquired from the same animal and the same set of electrodes by a commercial hardwired recording system to validate the wirel- ssly recorded signals.
收起
摘要 :
The amount of calibration data needed to produce images of adequate quality can prevent auto-calibrating parallel imaging reconstruction methods like generalized autocalibrating partially parallel acquisitions (GRAPPA) from achiev...
展开
The amount of calibration data needed to produce images of adequate quality can prevent auto-calibrating parallel imaging reconstruction methods like generalized autocalibrating partially parallel acquisitions (GRAPPA) from achieving a high total acceleration factor. To improve the quality of calibration when the number of auto-calibration signal (ACS) lines is restricted, we propose a sparsity-promoting regularized calibration method that finds a GRAPPA kernel consistent with the ACS fit equations that yields jointly sparse reconstructed coil channel images. Several experiments evaluate the performance of the proposed method relative to unregularized and existing regularized calibration methods for both low-quality and underdetermined fits from the ACS lines. These experiments demonstrate that the proposed method, like other regularization methods, is capable of mitigating noise amplification, and in addition, the proposed method is particularly effective at minimizing coherent aliasing artifacts caused by poor kernel calibration in real data. Using the proposed method, we can increase the total achievable acceleration while reducing degradation of the reconstructed image better than existing regularized calibration methods.
收起
摘要 :
This paper provides an alternative solution to the costly representation of multi-view video data, which can be used for both rendering and scene analyses. Initially, a new efficient Monte Carlo discrete surface reconstruction met...
展开
This paper provides an alternative solution to the costly representation of multi-view video data, which can be used for both rendering and scene analyses. Initially, a new efficient Monte Carlo discrete surface reconstruction method for foreground objects with static background is presented, which outperforms volumetric techniques and is suitable for GPU environments. Some extensions are also presented, which allow a speeding up of the reconstruction by exploiting multi-resolution and temporal correlations. Then, a fast meshing algorithm is applied, which allows interpolating a continuous surface from the discrete reconstructed points. As shown by the experimental results, the original video frames can be approximated with high accuracy by projecting the reconstructed foreground objects onto the original viewpoints. Furthermore, the reconstructed scene can be easily projected onto any desired virtual viewpoint, thus simplifying the design of free-viewpoint video applications. In our experimental results, we show that our techniques for reconstruction and meshing compare favorably with the state-of-the-art, and we also introduce a rule-of-thumb for effective application of the method with a good quality versus representation cost trade-off.
收起
摘要 :
Digital signal processors (DSPs) and field-programmable gate arrays (FPGAs) are predominant in the implementation of digital controllers and/or modulators for power converter applications. This paper presents a systematic comparis...
展开
Digital signal processors (DSPs) and field-programmable gate arrays (FPGAs) are predominant in the implementation of digital controllers and/or modulators for power converter applications. This paper presents a systematic comparison between these two technologies, depicting the main advantages and drawbacks of each one. Key programming and implementation aspects are addressed in order to give an overall idea of their most important features and allow the comparison between DSP and FPGA devices. A classical linear control strategy for a well-known voltage-source-converter (VSC)-based topology used as Static Compensator (STATCOM) is considered as a driving example to evaluate the performance of both approaches. A proof-of-concept laboratory prototype is separately controlled with the TMS320F2812 DSP and the Spartan-3 XCS1000 FPGA to illustrate the characteristics of both technologies. In the case of the DSP, a virtual floating-point library is used to accelerate the control routines compared to double precision arithmetic. On the other hand, two approaches are developed for the FPGA implementation, the first one reduces the hardware utilization and the second one reduces the computation time. Even though both boards can successfully control the STATCOM, results show that the FPGA achieves the best computation time thanks to the high degree of parallelism available on the device.
收起
摘要 :
This work addresses the problem of the optimal micro-siting of the wind turbines in large offshore wind power plants with the aim of maximizing the economic profitability of the project. To achieve this goal it is first necessary ...
展开
This work addresses the problem of the optimal micro-siting of the wind turbines in large offshore wind power plants with the aim of maximizing the economic profitability of the project. To achieve this goal it is first necessary to estimate the required investment and, secondly, the yearly operation and maintenance costs as well as the yearly income resulting from the operation of the wind power plant over its life span. With this purpose, a complete and realistic model of economic behavior for offshore wind farms has been developed.
收起
摘要 :
This paper presents the design, optimization, and experimental characterization of a planar antenna subarray for space applications. The proposed component is intended to be used as a building block for a large Ka-band antenna sys...
展开
This paper presents the design, optimization, and experimental characterization of a planar antenna subarray for space applications. The proposed component is intended to be used as a building block for a large Ka-band antenna systems. It is a one-dimensional linear array fed using a standard rectangular waveguide. It includes a transition between the waveguide and the beam forming network (BFN), the BFN itself and six radiators. A planar BFN and planar radiating aperture-fed double stacked circular patches are used to guarantee the low-profile of the proposed element. To ensure low losses and to reduce the mutual coupling inside the BFN, the suspended stripline (SSL) is selected as the transmission line. Due to the tight requirements on RF performance and the limited space available, novel components, such as waveguide to SSL transition and SSL power dividers, are proposed. The design is optimized using a step-by-step procedure. A complete subarray prototype has been fabricated requiring no-less than six different manufacturing technologies. The experimental results have validated the numerical simulations and show a good overall performance and low losses with efficiency above 70% in the frequency band 25.5–27 GHz (5.7%).
收起
摘要 :
We investigate the propagation of surface waves along a spatially dispersive graphene sheet, including substrate effects. The proposed analysis derives the admittances of an equivalent circuit of graphene able to handle spatial di...
展开
We investigate the propagation of surface waves along a spatially dispersive graphene sheet, including substrate effects. The proposed analysis derives the admittances of an equivalent circuit of graphene able to handle spatial dispersion, using a nonlocal model of graphene conductivity. Similar to frequency-selective surfaces, the analytical admittances depend on the propagation constant of the waves traveling along the sheet. Dispersion relations for the supported TE and TM modes are then obtained by applying a transverse resonance equation. Application of the method demonstrates that spatial dispersion can dramatically affect the propagation of surface plasmons, notably modifying their mode confinement and increasing losses, even at frequencies where intraband transitions are the dominant contribution to graphene conductivity. These results show the need to correctly assess spatial dispersion effects in the development of plasmonic devices at the low THz band.
收起
摘要 :
In this paper, the line-offset offset-open offset-short (LZZ) calibration technique for vector network analyzers (VNA) is introduced. The LZZ uses as calibration standards a fully known transmission line and two offset reflecting ...
展开
In this paper, the line-offset offset-open offset-short (LZZ) calibration technique for vector network analyzers (VNA) is introduced. The LZZ uses as calibration standards a fully known transmission line and two offset reflecting loads. The mathematical formulation of the LZZ is based on the use of $ABCD$-parameters for modeling the imperfect VNA as well as calibration standards. As a result, it is shown that the error coefficients characterizing the imperfect VNA can be calculated by comparing the estimated impedance of the two loads with the characteristic impedance of the transmission line. In order to validate the proposed method, the line–reflect–line (LRL) and the line–reflect–reflect–match (LRRM) calibration techniques are used. A high correlation between the $S$-parameters of a heterostructure field-effect transistor corrected with the LRL, LRRM, and LZZ techniques up to 45 GHz is achieved.
收起
