26 July 2018 in IEEE Transactions on Instrumentation and Measurement
While numerous works analyze the theoretical background of compressive sensing (CS) and provide rich mathematical theory, few practical implementations exist and they mostly share the same disadvantage of being either too complex or too expensive to implement. As a result of the reported work, a simple measurement setup for CS using off-the-shelf components is proposed. A simple and portable CS architecture consists of a digital camera, a projector, and mechanical integration elements. Extensive analysis of the proposed measurement system and the detailed calibration method is provided. The presented CS method allows for the use of a low-resolution camera with significant lens distortion for data acquisition. The results of our CS method are nondistorted, high-resolution images, which are obtained from the acquired measurement data using sparse optimization.
20 November 2017 in IEEE/ACM Transactions on Audio, Speech, and Language Processing
This paper presents an extensive statistical analysis of the acoustic startle response of two vocal parameters: fundamental frequency (F0) and root-mean-square energy (E), as well as of the orbicularis oculi (eyeblink) surface electromyography (sEMG). An experiment was conducted in which fourteen participants were exposed to acoustic startle stimuli of varying parameters, i.e., intensity level, duration, rise time, and spectral type, during periods of sustained phonation. Voice recordings of the phonations were taken alongside several physiological signals, of which only the sEMG was analyzed in this paper. Response features (peak value, peak time, latency, rise time, fall time, and duration) were extracted on F0, E and sEMG data, and statistical analysis was conducted using linear mixed effects models to show the response behavior with the varying stimuli. The results for vocal F0 and E data were congruent with sEMG data and earlier work in the field. The results demonstrated that vocal analysis can be used as a feasible alternative to the sEMG eyeblink analysis of acoustic startle responses.
18 January 2011 in Automatika
Currently, one of the most widely used distance measures in speech and speaker recognition is the Euclidean distance between mel frequency cepstral coefficients (MFCC). MFCCs are based on filter bank algorithm whose filters are equally spaced on a perceptually motivated mel frequency scale. The value of mel cepstral vector, as well as the properties of the corresponding cepstral distance, are determined by several parameters used in mel cepstral analysis. The aim of this work is to examine compatibility of MFCC measure with human perception for different values of parameters in the analysis. By analysing mel filter bank parameters it is found that filter bank with 24 bands, 220 mels bandwidth and band overlap coefficient equal and higher than one gives optimal spectral distortion (SD) distance measures. For this kind of mel filter bank, the difference between vowels can be recognised for full-length mel cepstral SD RMS measure higher than 0.4 - 0.5 dB. Further on, we will show that usage of truncated mel cepstral vector (12 coefficients) is justified for speech recognition, but may be arguable for speaker recognition. We also analysed the impact of aliasing in cepstral domain on cepstral distortion measures. The results showed high correlation of SD distances calculated from aperiodic and periodic mel cepstrum, leading to the conclusion that the impact of aliasing is generally minor. There are rare exceptions where aliasing is present, and these were also analysed.
15 April 2011 in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
We propose a method for direct digital frequency synthesis (DDS) using a cubic spline piecewise-polynomial model for a phase-to-sinusoid amplitude converter (PSAC). This method offers maximum smoothness of the output signal. Closed-form expressions for the cubic polynomial coefficients are derived in the spectral domain and the performance analysis of the model is given in the time and frequency domains. We derive the closed-form performance bounds of such DDS using conventional metrics: rms and maximum absolute errors (MAE) and maximum spurious free dynamic range (SFDR) measured in the discrete time domain. The main advantages of the proposed PSAC are its simplicity, analytical tractability, and inherent numerical stability for high table resolutions. Detailed guidelines for a fixed-point implementation are given, based on the algebraic analysis of all quantization effects. The results are verified on 81 PSAC configurations with the output resolutions from 5 to 41 bits by using a bit-exact simulation. The VHDL implementation of a high-accuracy DDS based on the proposed PSAC with 28-bit input phase word and 32-bit output value achieves SFDR of its digital output signal between 180 and 207 dB, with a signal-to-noise ratio of 192 dB. Its implementation requires only one 18 kB block RAM and three 18-bit embedded multipliers in a typical field-programmable gate array (FPGA) device.
28 January 2021 in IEEE Access
The precise stellar object identification is one of the major research fields in astronomy. In astronomical images, the 2D Gaussian function provides a good approximation of stellar objects. This paper proposes a robust method for the estimation of the 2D Gaussian profiles in astronomical images when the stellar components are overlapped. The proposed method efficiently separates the regions of interest of close components without knowing neither their exact number nor accurate center positions. Therefore, fainter objects can be identified, and outliers can be detected. The proposed method is an iteratively reweighted least squares method where weights are updated in each step using a robust measure of residual error dispersion. Soft thresholding based on robust statistics increases the accuracy of the estimated 2D profile parameters and efficiently solves the problem of outliers. The comparison of the proposed method with the conventional least squares method showed a significant modeling gain for one particular set of contributing factors; the difference of magnitudes of close objects, their center-to-center distance, and the widths of their semiaxes relative to the imaging system resolution.
5 October 2022 in IEEE Access
This paper estimates the multidimensional Gaussian profile parameters from the noisy measurements in the exponential function’s argument domain. The proposed method minimizes the weighted squared error between the natural logarithm of the model and the logarithm of the normalized input data with the weights set to the input data values or model values. The proposed method is an iterative method where the parameters of the covariance matrix and the profile’s peak position are alternatively estimated. The main advantage of the proposed method is a one-step analytical solution for the parameters of the covariance matrix and the linear profile scale for the given initial centroid position for arbitrary dimensions. The profile’s peak position is then updated given the estimated parameters by solving a system of nonlinear coupled equations using an iterative optimization procedure. Finally, the proposed method in the log domain is compared with the LS method in the domain of Gaussian profile values, where all profile parameters are simultaneously estimated using an iterative procedure for solving a system of nonlinear equations using numerical optimization. The proposed log domain estimation method yields similar results as the numerical LS method in the value domain for sufficiently high signal-to-noise ratios (SNRs) and narrow regions-of-interest (ROIs) concerning their precision. However, it converges much faster due to the analytic solution.