Filter Bank Parameters Research Articles

A Deep Dive into De-Doppler Algorithms for SETI

Narrowband SETI algorithms have been in use for many decades. A key processing step in the usual SETI pipeline called “De-Doppler” involves non-coherent integration of linearly-drifting lines in a spectrogram. De- Doppler (DD) is needed because a transmitted constant-frequency tone will be received as a chirp, or a tone with a near-constant frequency rate, due to relative accelerations between the transmitter and receiver. The frequency drift may be an important discriminator against certain types of radio-frequency interference. This paper examines the standard Taylor DD algorithm and finds there are significant signal losses in the integration process. Part of the loss is caused by imperfect indexing in the algorithm. Additional loss is associated with the effect of an FFT filter bank on chirp signals; DD performance is intertwined with the filter bank. Filter bank improvements involving polyphase filter banks are examined. A new DD algorithm called “fastDD” is proposed, as well as extensions to the Taylor algorithm. They both offer significant performance gains and higher resolution in frequency and drift rate in the DD detection plane. The loss mechanisms are now understood. Computational cost trade-offs for various filter bank and DD parameters are presented based on C code functions. Several “best” options are given which offer significant SNR improvements ranging from 2 to 4 dB for modest levels of additional computation. These options have the potential to dramatically improve detection rates.

Drone classification using RF signal based spectral features

Drone detection and classification, important in military and civilian applications, are performed using different sensor signals. Proposed study handles this task using Radio Frequency (RF) signals utilizing basic machine learning methods. It is composed of two main stages as feature extraction succeeded by training/testing of the model. In feature extraction stage, valuable information for classification, contained in the RF signal, is obtained. For this purpose, spectral features, frequently used in speech processing applications, are employed. Specifically, Power Spectral Density (PSD), Mel-Frequency Cepstral Coefficients (MFCC) and Linear Frequency Cepstral Coefficients (LFCC) are adopted by adjusting filter bank margins and parameters for this task. In the second stage, a Support Vector Machine (SVM) classifier is first trained based on the obtained features and finally tested to measure its performance. All experimental studies are carried out using publicly available DroneRF dataset. This dataset contains 2-Class, 4-Class and 10-Class samples for drone existence vs. background (BG), drone types and drone operation modes, respectively. The best classification results are obtained using, PSD, MFCC and LFCC based features for 2-Class, MFCC and LFCC based features for 4-Class and LFCC based features for 10-Class. Accuracy rates for 2-Class, 4-Class and 10-Class are 100%, 98.67% and 95.15%, respectively. These results show that the proposed method outperforms the results given in the literature for DroneRF dataset.

Deep Generalization of Structured Low-Rank Algorithms (Deep-SLR).

Structured low-rank (SLR) algorithms, which exploit annihilation relations between the Fourier samples of a signal resulting from different properties, is a powerful image reconstruction framework in several applications. This scheme relies on low-rank matrix completion to estimate the annihilation relations from the measurements. The main challenge with this strategy is the high computational complexity of matrix completion. We introduce a deep learning (DL) approach to significantly reduce the computational complexity. Specifically, we use a convolutional neural network (CNN)-based filterbank that is trained to estimate the annihilation relations from imperfect (under-sampled and noisy) k-space measurements of Magnetic Resonance Imaging (MRI). The main reason for the computational efficiency is the pre-learning of the parameters of the non-linear CNN from exemplar data, compared to SLR schemes that learn the linear filterbank parameters from the dataset itself. Experimental comparisons show that the proposed scheme can enable calibration-less parallel MRI; it can offer performance similar to SLR schemes while reducing the runtime by around three orders of magnitude. Unlike pre-calibrated and self-calibrated approaches, the proposed uncalibrated approach is insensitive to motion errors and affords higher acceleration. The proposed scheme also incorporates image domain priors that are complementary, thus significantly improving the performance over that of SLR schemes.

Learning Front-end Filter-bank Parameters using Convolutional Neural Networks for Abnormal Heart Sound Detection.

Automatic heart sound abnormality detection can play a vital role in the early diagnosis of heart diseases, particularly in low-resource settings. The state-of-the-art algorithms for this task utilize a set of Finite Impulse Response (FIR) band-pass filters as a front-end followed by a Convolutional Neural Network (CNN) model. In this work, we propound a novel CNN architecture that integrates the front-end band-pass filters within the network using time-convolution (tConv) layers, which enables the FIR filter-bank parameters to become learnable. Different initialization strategies for the learnable filters, including random parameters and a set of predefined FIR filter-bank coefficients, are examined. Using the proposed tConv layers, we add constraints to the learnable FIR filters to ensure linear and zero phase responses. Experimental evaluations are performed on a balanced 4-fold cross-validation task prepared using the PhysioNet/CinC 2016 dataset. Results demonstrate that the proposed models yield superior performance compared to the state-of-the-art system, while the linear phase FIR filter-bank method provides an absolute improvement of 9.54% over the baseline in terms of an overall accuracy metric.

Split-spectrum processing technique for SNR enhancement of ultrasonic guided wave

Ultrasonic guided wave (UGW) systems are broadly used in several branches of industry where the structural integrity is of concern. In those systems, signal interpretation can often be challenging due to the multi-modal and dispersive propagation of UGWs. This results in degradation of the signals in terms of signal-to-noise ratio (SNR) and spatial resolution. This paper employs the split-spectrum processing (SSP) technique in order to enhance the SNR and spatial resolution of UGW signals using the optimized filter bank parameters in real time scenario for pipe inspection. SSP technique has already been developed for other applications such as conventional ultrasonic testing for SNR enhancement. In this work, an investigation is provided to clarify the sensitivity of SSP performance to the filter bank parameter values for UGWs such as processing bandwidth, filter bandwidth, filter separation and a number of filters. As a result, the optimum values are estimated to significantly improve the SNR and spatial resolution of UGWs. The proposed method is synthetically and experimentally compared with conventional approaches employing different SSP recombination algorithms. The Polarity Thresholding (PT) and PT with Minimization (PTM) algorithms were found to be the best recombination algorithms. They substantially improved the SNR up to 36.9dB and 38.9dB respectively. The outcome of the work presented in this paper paves the way to enhance the reliability of UGW inspections.

Perceptual Significance of Cepstral Distortion Measures in Digital Speech Processing

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.

An efficient gbest-guided Cuckoo Search algorithm for higher order two channel filter bank design

Abstract This paper proposes a new algorithm based on Gbest-guided Cuckoo Search (GCS) algorithm for the design of higher order Quadrature Mirror Filter (QMF) bank. Although the optimization of lower order filters can be performed easily by applying existing meta-heuristic optimization techniques like Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC) etc., these methods are unsuccessful in searching higher order filter coefficients due to multimodality and nonlinear problem space; leads to some undesirable behaviors in filter responses like ripples in transition band, lower stop-band attenuation etc.. Comparison with other available results in the literature indicate that the proposed method exhibits an 69.02% increase in stop-band attenuation and 99.71% reduction in Perfect Reconstruction Error (PRE) of higher order filter bank. Besides, the percentage improvements in Fitness Function Evaluations (FFEs) of GCS based 55th order QMF bank design with respect to PSO, ABC and CSA are 81%, 82% and 59% respectively, and execution time is improved by 73%, 72% and 42% respectively. The simulation results also reveal that the proposed approach exhibits lowest mean and variance in different assessment parameters of filter bank and it does not require tuning of algorithmic parameters whereas in standard CSA replacement factor need to be adjusted. Further, the proposed algorithm is tested on six standard benchmark problems and complex benchmark functions from the CEC 2013 where it demonstrated significant performance improvements than other existing methods.

Speech frame recognition based on less shift sensitive wavelet filter banks

The wavelet transform possesses multi-resolution property and high localization performance; hence, it can be optimized for speech recognition. In our previous work, we show that redundant wavelet filter bank parameters work better in speech recognition task, because they are much less shift sensitive than those of critically sampled discrete wavelet transform (DWT). In this paper, three types of wavelet representations are introduced, including features based on dual-tree complex wavelet transform (DT-CWT), perceptual dual-tree complex wavelet transform, and four-channel double-density discrete wavelet transform (FCDDDWT). Then, appropriate filter values for DT-CWT and FCDDDWT are proposed. The performances of the proposed wavelet representations are compared in a phoneme recognition task using special form of the time-delay neural networks. Performance evaluations confirm that dual-tree complex wavelet filter banks outperform conventional DWT in speech recognition systems. The proposed perceptual dual-tree complex wavelet filter bank results in up to approximately 9.82 % recognition rate increase, compared to the critically sampled two-channel wavelet filter bank.

Constructing totally interpolating wavelet frame systems

The system of totally interpolating wavelet frames is discussed in this paper, in which both the scaling function and one of wavelet functions are interpolating. It will be shown that corresponding filter banks possess the special structure, and the parametrization of filter banks is present. Moreover, we show that when considering tight frame systems with two generators, the Ron–Shen's continuous-linear-spline-based tight frame is the only one with totally interpolating property and symmetry. But in the dual frame context, more good examples of bi-frames with symmetric/antisymmetric property can be obtained and constructed, which in particular, include frames with the uniform symmetry.

Cochlear Implant Filterbank Design and Optimization: A Simulation Study

Cochlear implants (CIs) are devices capable of restoring hearing function in profoundly-deaf patients to an acceptable degree of performance. An essential processing step in any cochlear implant is frequency analysis, which is usually performed via banks of filters. Here, we simulate and test the suitability of different filters and filterbank architectures for CIs with respect to their performance in speech intelligibility. Four different filters were implemented in an established model of CI hearing, the tone-excited vocoder, namely: GTF (Gammatone Filter), DAPGF (Differentiated All-Pole GTF), OZGF (One-Zero GTF) and BUTF (Butterworth). Three filterbank parameters, the filter order ( N), the filter quality factor ( Q) and the number of channels ( Ch), and their combinations were tested using objective and subjective metrics. Simulation results show that all filters tested are suitable for CI implementation, but that the choice of Q and N parameter values is crucial. For most conditions, optimal ( N,Q) combinations were within few units away from the combination (2, 4).

A Low-Power and High-Precision Programmable Analog Filter Bank

Analog filter banks befit remote audio- and vibration-sensing applications, which require frequency analysis to be performed with low-power consumption and with moderate-to-high precision. The precision of a filter bank depends on both the signal-path precision (i.e., dynamic range) and also the parameter precision (e.g., accuracy of the center frequencies). This brief presents a new bandpass filter for audio-frequency filter banks and provides a procedure for designing this filter. The filter is used in a 16-channel filter bank which has been fabricated in a 0.35- CMOS process. This filter bank has a dynamic range exceeding 62 dB and consumes only 63.6 when biased for speech frequencies. The filter bank's parameters are set via floating-gate current sources. This brief shows how to use these floating gates to obtain a versatile filter bank that can be precisely reprogrammed to arbitrary filter spacings and frequency weightings, with a parameter accuracy exceeding 99%.

Hybrid Filterbank ADCs With Blind Filterbank Estimation

The hybrid filterbank architecture permits implementing accurate, high speed analog-to-digital converters. However, its design requires an accurate knowledge of the analog filterbank parameters, which is difficult to have due to the nonstationary nature of these parameters. This paper proposes a blind estimation method for the analog filterbank parameters, which is able to cope with nonstationary input signals. This is achieved by using the notion of averaged input spectrum. The estimated parameters are used to reconstruct the samples in a least mean squares (LMS) sense. The proposed LMS design generalizes existing approaches by dropping the bandlimited assumption on the input signal. Instead, it assumes that the input has an arbitrary power spectrum which is adaptively estimated. Numerical experiments are presented showing the good performance of the blind estimation stage and the clear advantage of the proposed LMS design.

Support vector wavelet adaptation for pathological voice assessment

The presence of abnormalities in the vocal system affects the quality of the voice and changes its characteristics. Digital analysis of pathological voices can be an effective and non-invasive tool for the detection of such alterations. This paper proposes a wavelet-based method to distinguish between normal and disordered voices. Wavelet filter banks are used in conjunction with support vector machines, as feature extractors and classifiers, respectively. Orthogonal filter banks are implemented using a highly efficient structure known as “lattice” that parameterizes filter banks and produces a few parameters. The overall problem is to find these parameters such that perfect classification is achieved. To search for such parameters, a genetic algorithm with a fitness function corresponding to the classification result is applied. Simulation is done on the KAY database (a comprehensive database including 710 normal and pathological voice signals, developed by the Massachusetts Eye and Ear Infirmary Voice and Speech Lab), and one additional test set. It is observed that a genetic algorithm is able to find the filter bank parameters such that a 100% correct classification rate is achieved in classifying normal and pathological voices when the test is performed on both databases.

Fabric Defect Detection Method Based on Gabor Filters

In this paper, a fabric defect detection method based on Gabor filter bank is present. In this method, the parameters of Gabor filter bank depend on the fabric texture feature. Using the Gabor filter bank with multi-frequency and multi-orientation, a textile image produces multi-image. Then, the images are reconstructed into one image for detecting defects. It is illustrated that most kinds of defects are correctly detected and segmented. The experimental results show that the algorithm is robust and has good detection effect.

Perceptual Significance of Cepstral Distortion Measures in Digital Speech Processing

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.

Real-world acoustic event detection

Acoustic Event Detection (AED) aims to identify both timestamps and types of events in an audio stream. This becomes very challenging when going beyond restricted highlight events and well controlled recordings. We propose extracting discriminative features for AED using a boosting approach, which outperform classical speech perceptual features, such as Mel-frequency Cepstral Coefficients and log frequency filterbank parameters. We propose leveraging statistical models better fitting the task. First, a tandem connectionist-HMM approach combines the sequence modeling capabilities of the HMM with the high-accuracy context-dependent discriminative capabilities of an artificial neural network trained using the minimum cross entropy criterion. Second, an SVM–GMM-supervector approach uses noise-adaptive kernels better approximating the KL divergence between feature distributions in different audio segments. Experiments on the CLEAR 2007 AED Evaluation set-up demonstrate that the presented features and models lead to over 45% relative performance improvement, and also outperform the best system in the CLEAR AED Evaluation, on detection of twelve general acoustic events in a real seminar environment.

Construction of parametric biorthogonal wavelet filter banks with two parameters for image coding

We had presented a simple technique, which is based on the theory of Diophantine equation, for parametrization of popular biorthogonal wavelet filter banks (BWFBs) having the linear phase and arbitrary multiplicity of vanishing moments (VMs), and constructed a type of parametric BWFBs with one free parameter [15]. Here we generalize this technique to the case of two parameters, and construct a type of parametric BWFBs with two free parameters. The closed-form parameter expressions of the BWFBs are derived, with which any two-parameter family of BWFBs having preassigned VMs can be constructed, and six families, i.e., 9/11, 10/10, 13/11, 10/14, 17/11, and 10/18 families, are considered here. Two parameters provide two degrees of freedom to optimize the resulting BWFBs with respect to other criteria. In particular, in each family, three specific rational-coefficient BWFBs with attractive features are obtained by adjusting the parameters: the first is not only very close to a quadrature mirror filter (QMF) bank, but has optimum coding gain; the second possesses characteristics that are close to the irrational BWFB with maximum VMs by Cohen et al.; and the last which has binary coefficients can realize a multiplication-free discrete wavelet transform. In addition, two BWFBs are systematically verified to exhibit performance competitive to several state-of-the-art BWFBs for image compression, and yet require lower computational costs.

Optimization Design of Biorthogonal Wavelets for Embedded Image Coding

We present here a simple technique for parametrization of popular biorthogonal wavelet filter banks (BWFBs) having vanishing moments (VMs) of arbitrary multiplicity. Given a prime wavelet filter with VMs of arbitrary multiplicity, after formulating it as a trigonometric polynomial depending on two free parameters, we prove the existence of its dual filter based on the theory of Diophantine equation. The dual filter permits perfect reconstruction (PR) and also has VMs of arbitrary multiplicity. We then give the complete construction of two-parameter families of 17/11 and 10/18 BWFBs, from which any linear-phase 17/11 and 10/18 BWFB possessing desired features could be derived with ease by adjusting the free parameters. In particular, two previously unpublished BWFBs for embedded image coding are constructed, both have optimum coding gains and rational coefficients. Extensive experiments show that our new BWFBs exhibit performance equal to Winger's W--17/11 and Villasenor's V--10/18 (superior to CDF--9/7 by Cohen et al. and Villasenor's V--6/10) for image compression, and yet require slightly lower computational costs.

Design of Optimal Quincunx Filter Banks for Image Coding

Two new optimization-based methods are proposed for the design of high-performance quincunx filter banks for the application of image coding. These new techniques are used to build linear-phase finite-length-impulse-response (FIR) perfect-reconstruction (PR) systems with high coding gain, good frequency selectivity, and certain prescribed vanishing-moment properties. A parametrization of quincunx filter banks based on the lifting framework is employed to structurally impose the PR and linear-phase conditions. Then, the coding gain is maximized subject to a set of constraints on vanishing moments and frequency selectivity. Examples of filter banks designed using the newly proposed methods are presented and shown to be highly effective for image coding. In particular, our new optimal designs are shown to outperform three previously proposed quincunx filter banks in 72% to 95% of our experimental test cases. Moreover, in some limited cases, our optimal designs are even able to outperform the well-known (separable) 9/7 filter bank (from the JPEG-2000 standard).

Extracting tissue deformation using Gabor filter banks.

This paper presents a new approach for accurate extraction of tissue deformation imaged with tagged MR. Our method, based on banks of Gabor filters, adjusts (1) the aspect and (2) orientation of the filter's envelope and adjusts (3) the radial frequency and (4) angle of the filter's sinusoidal grating to extract information about the deformation of tissue. The method accurately extracts tag line spacing, orientation, displacement and effective contrast. Existing, non-adaptive methods often fail to recover useful displacement information in the proximity of tissue boundaries while our method works in the proximity of the boundaries. We also present an interpolation method to recover all tag information at a finer resolution than the filter bank parameters. Results are shown on simulated images of translating and contracting tissue.