Showing papers on "Quadrature mirror filter published in 1999"

Modified DFT filter banks with perfect reconstruction

Abstract: In this paper, essential features of the recently introduced modified discrete Fourier transform DFT (MDFT) filter bank are presented. First, it is shown that all analysis and synthesis filters-obtained by appropriate complex modulation of a low-pass prototype filter-are linear phase. This is important for subband image coding applications. Another important property is the structure-inherent alias cancellation: all odd alias spectra are automatically compensated in the synthesis filter bank. Further, the MDFT filter bank provides perfect reconstruction for the same prototypes as for cosine-modulated filter banks. Thus, the same design methods can be used. Finally, different mappings of the input signal into the subbands are discussed and a comparison to the well-known cosine-modulated filter banks is given.

A new directional filter bank for image analysis and classification

Abstract: A new directional filter bank for image analysis and classification is proposed. This paper introduces an improved structure in order to visualize subband outputs of the directional filter banks, while retaining the attractive properties of the original directional filter banks such as 1-D separable filtering, perfect reconstruction, and maximal decimation. Using this structure, any arbitrary 2/sup n/ band directional filter bank can be implemented by cascading simple directional filter bank blocks, unlike the original structure that needs a parallel structure for visualizing subband outputs. Also, in order to have nondistorted phase information in the subbands for visualization, both FIR and IIR filter prototypes that can be implemented efficiently are provided for linear phase filtering. This paper shows the approach proposed here can be applied to image analysis and classification.

Extending the threshold and frequency range for phase-based frequency estimation

Abstract: Standard phase-based frequency estimation has a threshold that is high, is frequency dependent, and does not decrease with increasing signal length. These problems are solved by processing the signal with a highly overlapped filter bank before applying phase-based frequency estimation. By exploiting decimation, a closed-form matrix inversion, and cascades of simple filter banks, the computational complexity is kept low.

Direct conversion receiver with wide band analog frequency conversion front end and digital demodulating and selecting back end

Abstract: A radio-frequency signal receive section, including an antenna, collectively receives radio-frequency signals on a plurality of channels each of which is modulated with data and outputs a received signal. In a quadrature demodulator, the received signal is multiplied by each of first quadrature local signals to produce quadrature baseband signals. In two low-pass filters, low-frequency component signals are extracted from the quadrature baseband signals. In a quadrature modulator, the low-frequency component signals are multiplied by second quadrature local signals and thereby converted into quadrature intermediate-frequency signals, which are in turn combined into an output signal. After converted into a digital signal, the output signal of the quadrature modulator is inputted to a digital signal processing unit where a signal on at least one channel is selected and the selected signal is processed to recover the original data.

Radar receiver having matched filter processing

Abstract: An electronic counter-measure receiver can be colocated with a threat simulator radar, which transmits a radar signal. As a pulsed-Doppler signal, this transmitted radar signal is sampled as a reference signal by the receiver to determine quadrature, coherent matched filter coefficients. A return signal is received and processed with the reference signal as quadrature signal components with the quadrature, coherent matched filter coefficients within a matched filter. The return signal is processed as a first intermediate frequency via a mixer and local oscillator and a second quadrature intermediate frequency via a digital local oscillator and mixer. A fast Fourier transform is performed to determine Doppler information. In a continuous wave transmission for sampling arbitrary but unevenly spaced in time pulses are generated from a pulse generator source internal to the receiver.

An efficient algorithm to design weighted minimax perfect reconstruction quadrature mirror filter banks

Abstract: An efficient algorithm is presented to design lattice structure two-channel perfect reconstruction quadrature mirror filter (PR-QMF) banks. We formulate the filter bank design problem as an unconstrained weighted least squares problem with respect to the coefficients of the lattice structure. The proposed iterative algorithm optimizes the lattice coefficients and provides flexible control of the filters' stopband ripple profiles. Typically, only a few iterations of the algorithm are needed to obtain an optimal solution in the weighted minimax sense. We include a set of practical design rules for use with our algorithm. These rules allow very good estimation of important filter bank characteristics, such as the filter length and the number of signed digits for quantization of the lattice coefficients into canonic signed digit representation, to meet a given set of PR-QMF bank specifications.

Improved weighted least squares algorithm for the design of quadrature mirror filters

Abstract: This paper presents an improved weighted least squares (WLS) algorithm for the design of quadrature mirror filters (QMFs), First, a new term is incorporated into the objective function that effectively prevents an optimization algorithm from producing suboptimal QMFs. These suboptimal QMFs exhibit a transition band anomaly; the frequency responses of the filters have large oscillatory components in the transition band. The new term can be applied to the WLS design of any FIR filter to prevent a similar transition band anomaly. Next, we present an algorithm to obtain the QMF coefficients that minimize the objective function incorporating the new term. The computational requirement of this algorithm is also briefly discussed. Last, we include a set of practical design rules for use with our algorithm. These rules simplify the design process by providing good estimation of the design parameters, such as the minimum filter length, to meet a given set of QMF specifications.

Discrete Lagrangian methods for optimizing the design of multiplierless QMF banks

Abstract: In this paper, we present a new discrete Lagrangian method for designing multiplierless quadrature mirror filter banks. The filter coefficients in these filter banks are in powers-of-two, where numbers are represented as sums or differences of powers of two (also called canonical signed digit representation), and multiplications are carried out as additions, subtractions, and shifts. We formulate the design problem as a nonlinear discrete constrained optimization problem, using reconstruction error as the objective, and stopband and passband energies, stopband and passband ripples, and transition bandwidth as constraints. Using the performance of the best existing designs as constraints, we search for designs that improve over the best existing designs with respect to all the performance metrics. We propose a new discrete Lagrangian method for finding good designs and study methods to improve the convergence speed of Lagrangian methods without affecting their solution quality. This is done by adjusting dynamically the relative weights between the objective and the Lagrangian part. We show that our method can find designs that improve over Johnston's benchmark designs using a maximum of three to six ONE bits in each filter coefficient instead of using floating-point representations. Our approach is general and is applicable to the design of other types of multiplierless filter banks.

New filter banks and more regular wavelets

Abstract: One of the most interesting features of a wavelet is its Sobolev regularity. In this paper, we construct new wavelets that are more regular than the Daubechies wavelets for a given support width. We tabulate the coefficients of the new filters to make them easily accessible. We show that these filters outperform the Daubechies filters in the L/sup 2/ approximation of the ideal filter. An application for speech analysis, synthesis, and compression is provided.

Minimum mean square error FIR filter banks with arbitrary filter lengths

Abstract: The performance of subband image coders depends on the proper choice of filter banks. It is therefore desirable to find algorithms for optimizing the filter banks. A theory for finding jointly optimized analysis and synthesis filter banks with arbitrary filter lengths and arbitrary delay through the filter bank is proposed. Theoretical results are included. Comparisons are made against the well known 5/3, 9/7, and 10/18 wavelets. The proposed filter banks, which are both rare and source dependent, have a better distortion rate performance for all cases.

Image coding using wavelets based on two-channel linear phase orthogonal IIR filter banks

Abstract: In this paper, we focus on two-channel perfect reconstruction IIR filter banks. Often two-channel FIR filter banks with additional characteristics such as the perfect reconstruction property and linear phase are used in subband systems. This type of filter banks belongs to the class called biorthogonal filter banks. While several works have pointed out that the perfect reconstruction IIR filter banks have both linear phase and orthogonality, the linear phase orthogonal IIR filter banks have never been applied to image coding. Then we apply the two-channel perfect reconstruction IIR filter banks consisted of allpass filters, with both linear phase and orthogonality, to image coding and show that the proposed IIR filter bank with 3 zeros at z=1 for the analysis lowpass filter, has less computational complexity and produces better results in terms of PSNR than the FIR filter bank, namely the two-channel linear phase biorthogonal (9,7)-taps. Consequently both linear phase and orthogonality are essential to improve coding performance.

Design of multichannel QMF banks via frequency-domain optimizations

Abstract: Design of uniform multirate filter banks is studied via frequency domain optimizations. Efficient and reliable design algorithms are developed using state-space computations. In our approach, analysis filter banks are designed to achieve frequency domain specifications dictated by subband coding requirements, and synthesis filter banks are designed to minimize the reconstruction error in frequency domain under the constraint of zero-aliasing error. The design criterion is chosen to be the /spl Hscr//sub /spl infin// or Chebyshev norm. A state-space solution is derived for /spl Hscr//sub /spl infin//, optimization, and numerical algorithms are developed to obtain the optimal-synthesis filter bank. Moreover, the asymptotic perfect reconstruction property (in the sense that time-delay approaches to infinity) is established for the optimal /spl Hscr//sub /spl infin//, solution of the synthesis filter bank. The results in this paper generalize our earlier work for the two channel case to the M-channel case.

Subband decomposition based image compression algorithms with nonlinear adaptive filter banks.

Abstract: Figure 6: 128 128 windows from Barbara images a.) original and decoded at 0:5bpp by b.)median c.)Ad-lin d.)Ad-FFT-L e.)Ad-hybrid filters. Figure 7: 100 100 windows from Cameraman images a.) original and decoded at 0:5bpp by b.)median c.)Ad-FFT-L d.)Ad-hybrid filters. 6. CONCLUSIONS Image compression via subband decomposition is studied. An M-channel subband decomposition is obtained by using MSE-adaptive perfect reconstruction filter banks. The use of FFT-ordered L-filters which have MSE-adaptive time varying coefficients is proposed. L-filters include different linear and nonlinear filters and FFT-ordering structure proposes a unifying structure for different filtering operations. It is seen that (partially ordered) median filter performs better around edges whereas linear filter performs better in texture regions and FFT-L-filter performs better elsewhere. A texture detection algorithm which works over the already transmitted subbands is used to decide where to switch between different filters. The resulting algorithm outperforms both visually and quantitatively for different types of images having different characteristics. [4] ¨ O. N. Gerek and A. Enis C ¸ etin, " Adaptive Polyphase Subband Decomposition Structures for Image Compression , " presented at ICASSP'98 and submitted to IEEE Trans. on Image Processing. Figure 1: Analysis and synthesis stages of the 2-channel adaptive filter bank structure. 2 2 P + + 2 2-1 z x x h l P z-1 x ^ x x x 2 1-+

Processing finite length signals via filter banks without border distortions: a non-expansionist solution

Abstract: In this paper we introduce a novel and general matrix formulation of classical signal extension methods for subband processing of finite length signals. Considering a paraunitary 2-channel filter bank as a transformation cell, this new characterization makes it possible to show that perfect reconstruction of finite signals can be ensured without resorting to extra subband samples; thus, by using some traditional signal extension methods, non-expansionist transforms can be defined. Some of these transformations are analyzed to illustrate our theoretical results.

Design of perfect reconstruction integer-modulated filter banks

Abstract: New methods for the design of integer-modulated filter banks are presented. The novelty lies in the design of the modulation vectors, which are not restricted to certain structures like dyadic symmetry. In addition, an efficient implementation based on Householder factorization is presented. In all cases considered, both the filter prototype and the modulation sequences are composed of integers while maintaining perfect reconstruction for arbitrary input signals.

New results in multidimensional linear phase filter bank design

Abstract: We consider the multidimensional version of the problem of linear phase (LP) perfect reconstruction (PR) filter bank design. The filter bank design problem is posed as a matrix completion problem in the context of polynomial matrices having certain symmetries dictated by the linear phase property of the filter bank. We examine the usefulness of a strategy that succeeds in characterization and design of 1D linear phase filter banks via 2D examples. In the 2D quincunx case, for filter banks obtained by McClellan transformations a complete solution can be obtained via this method. The usefulness of the method remains questionable in more general situations. We discuss the issue with examples.

Error compensator for converting high frequency signals to base band feeds digitized signals of in-phase, quadrature branches to 2-channel adaptive filter, injects test signal during signal pauses

Abstract: The arrangement digitizes the HF signal or a derived IF signal after separation into in-phase and quadrature components. The digitized signals of the in-phase and quadrature branches are fed to a two-channel adaptive filter (20) and a test signal (5) is fed in before signal separation during signal pauses in which there is no signal for conversion.

Design of selective M-channel perfect reconstruction FIR filter banks

Abstract: A new method is presented for the optimal design of M-channel perfect reconstruction FIR quadrature mirror filter (QMF) banks which is based on the orthogonal matrix concept. The parameters of the analysis filters are adjusted so that they have a linear phase and minimum stopband energy. Comparison with a recently published method reveals that the resulting design procedure gives better stopband characteristics. A numerical design example is included to validate the theory and to demonstrate the superiority of the designed filters from selectivity, reconstruction and distortion points of view.

Novel FIR filter designs for digital quadrature demodulation

Abstract: Finite impulse response (FIR) filters can provide ideal group delay characteristics. This makes them attractive for implementing quadrature demodulators based on quadrature mixing and lowpass filtering concepts. This paper describes novel quadrature demodulators based on third-band FIR filters in which nearly a third of the filter coefficients are zero valued. The proposed designs can achieve high accuracy while retaining stop bands which are useful for suppressing undesired signals.

Multiresolutional modification of speech signals for listeners with hearing impairment.

Abstract: A new method, the multiresolutional modification algorithm (MMA), which modifies speech signals in frequency and time domains, is proposed for listeners with hearing impairment and early clinical results are reported. Unlike other methods, this algorithm modifies the wavelet coefficients of the speech signal in order to obtain a modified version of the original signal instead of modifying the speech waveform itself. The speech signal is first divided into subbands using an 11-level quadrature mirror filter (QMF) bank. These subbands are then modified using the modification algorithm. Finally, the inverse Wavelet transform is applied to these modified subband coefficients in order to reconstruct a modified version of the input signal. The efficacy of the MMA was evaluated using subjects with hearing impairment and subjects with no such impairment. Listening tests showed that the proposed algorithm increases the quality and intelligibility of the modified speech over the well-known modification algorithms.

An efficient top-down approach for the design of tree-structured orthonormal filter banks

Abstract: Wavelet and tree-structured filter bank based coding schemes find applications in a number of image coding algorithms. The efficiency of such coding schemes can be enhanced further by adapting the parameters of the coder to the statistics of the signal being compressed. We propose an algorithm to determine the coefficients of the filter bank matched to the signal statistics, at every node of a tree-structured filter bank. In the proposed algorithm, we visit each node in a top-down fashion and determine the bit-allocation and the coefficients of the filters used to carry out the decomposition. An estimate of the coding gain provided by the later stages is used to account for the interaction between the filter bank at the current node and the filter banks at the later nodes.

On the Application of the Parks-McClellan Algorithm to the Design of Quadrature Demodulation Filters

Abstract: : New results are presented concerning the use of the Parks-McClellan algorithm to design filters for digital quadrature demodulators based on quadrature mixing and lowpass filtering concepts. The use of a 4:1 ratio between the sampling rate and intermediate frequency to reduce computational cost complicates this problem. Since the in-phase (I) and quadrature (Q) filters become odd and even length filters, respectively, the matching of the passband gains becomes an important error source. Consequently, the problem is to find the best design for a pair of filters rather than the best design for a single filter. One issue is whether to design the I and Q filters separately, or derive them from a prototype filter. Another concerns techniques for designing fractional-band filters if these are desired. The performance data presented in this paper shows that the quadrature demodulator accuracy has a complex dependence on the approach and specifications used to design the filters. Since good matching of the filter gains in the passband occurred only under certain conditions, significant performance losses can occur unless some care is taken in designing the filters.

Eigenfilter design of real and complex coefficient prototypes for uniform and nonuniform filter banks

Abstract: In this study we propose a new method to design pseudo-QMF prototype filters to implement near perfect reconstruction (NPR) cosine-modulated filter banks. The method is based on the eigenfilter approach, that is simple to implement, but, nevertheless, is very efficient in designing high attenuation prototype filters. The method also allows to design complex coefficient prototypes that can be used to build nonuniform filter banks. The effectiveness of the method is demonstrated by means of some examples of design of both uniform and nonuniform filter banks.

Filter device for demodulated quadrature amplitude modulated signal

Abstract: The filter device has a filter circuit (10') with respective channels (17I,17Q) for the cosine and sine demodulated components of the quadrature amplitude modulated signal, with respective filter transmission functions for the demodulated components. A switching device (19,21) connects a transverse circuit path (9) between the input of one channel and the output of the other channel in one switch position and between the input of the second channel and the output of the first channel in a second switch position.

Complexity evaluation of conjugate quadrature mirror filter banks

Abstract: The aim of this paper is to assess the complexity of different implementations of CQF (Conjugate Quadrature Mirror Filters) analysis and synthesis filter banks. In particular, we compare the classical structure based on FIR (Finite Impulse Response) filters with the alternative implementation based on binomial filters. In both cases, we describe how it is possible to reduce complexity exploiting some inherent redundancies in the analysis or synthesis filter bank. Results show that, with respect to the FIR usual scheme, a complexity reduction of at least a factor 4 is achievable by a joint use of binomial filters and "smart" structures.

On the design of perfect reconstruction multiwavelet filter banks.

Abstract: The lifting scheme is a flexible method of designing perfect reconstruction multiwavelet filter banks for signal and image processing. This paper is on the usage of the lifting scheme in constructing such filter banks. It is shown that the constructed filter banks can be implemented very efficiently. Two algorithms on the design of multiwavelet filter banks using the lifting scheme are proposed. One is for the optimization of desirable properties based on the application concerned, and the other is for the reduction of computational complexity in implementation. Two examples are provided to illustrate the main ideas of the paper.

On EZW encoding of surveillance imagery matched for spatial scale on viewer resolution

Abstract: Embedded zerotree wavelet (EZW) image coding is efficient and uses significance ordering for coefficients. EZW encoding can be arbitrarily terminated resulting in high compression ratios with satisfactory image quality. Given a requested scale for an image section matched to the resolution of the viewer, when can the EZW encoding be terminated and how can a specified quality for the image representation be obtained? One can compute the effect from the aliasing of higher frequencies. The aliased signals are in the cut-off band of the analysis quadrature mirror filter. The more significant, and therefore earlier, EZW coefficients produce the majority of the alias error. So the encoding can be terminated once a specified residue alias error is reached. Then the update of the full precision of the relevant discrete wavelet transform coefficients for that scale are required to provide image representation on the viewer with low or no distortion.

Design of QMF banks based on derivative information

Abstract: New L2 objective functions for the design of quadrature mirror filter (QMF) banks are proposed. They are based on the derivative information of the reconstruction error. Simple and explicit matrix-form formulas for the proposed objective functions are derived. Efficient design methods are proposed by incorporating a separability technique into the derived optimality conditions on prototype filters. The proposed design methods need only solve linear equation iteratively without nonlinear optimisation. Design examples demonstrate that good low-delay QMF banks and linear-phase QMF banks can be obtained in only a few iterations. Compared with the conventional approach, the new approach leads to QMF banks with larger stopband attenuation and smaller reconstruction errors.