Journal Publications

• Anamitra Makur^†, Arigovindan Muthuvel^‡, P. Viswanadha Reddy^§, Eigenstructure approach for complete characterization of linear phase FIR perfect reconstruction analysis length 2M filterbanks, IEEE Transactions of Signal Processing. to be published

^†Nanyang Technological University, Singapore
^‡Swiss Federal Institute of Technology, Lausanne, Switzerland
^§General Electric, Bangalore, India

Abstract : The eigenstructure based characterization of M channel FIR PR filterbanks of [1] is extended here to the linear phase case. Some results relating to linear phase filterbanks is derived by finding appropriate restrictions on the Eigenstructure of the analysis polyphase matrix. Consequently a complete and minimal characterization for such filter banks with all analysis length 2M and any synthesis length is developed. Parameterization and design examples are also presented.

[1] A. Muthuvel and A. Makur, Eigenstructure approach for characterization of FIR PR filter banks with order one polyphase, IEEE Trans. Signal Processing, vol. 49, pp. 2283–2291, Oct. 2001.
 
 

• P. Viswanadha Reddy, Anamitra Makur, M Channel Causal Stable IIR Perfect Reconstruction Filter Bank Design Using Minimal Factorization, IEEE Transactions on Signal Processing. in review

Abstract : In this work, the problem of design ofM channel IIR causal stable perfect reconstruction filter banks are addressed and some designs based on the minimal factorization of the state space form of the polyphase matrix are presented based on the concept of factorization of rational matrix functions. The analysis and synthesis polyphase matrices are constrained to be minimal. In the first method, eigen values of the state transition matrices of the analysis and the synthesis bank may be different. In the second method, these eigen values are chosen to be same. A design technique to reduce the optimization cost based on sequential (stage-by-stage) design is also presented to reduce the optimization cost. Simulation results are also presented.
 
 
• P. Viswanadha Reddy, Anamitra Makur, M Channel Causal Stable IIR Perfect Reconstruction Filter Bank Design Using Jordan Decomposition, IEEE Transactions on Circuits and Systems. in review

Abstract : In this work, the problem of design of M channel IIR causal stable perfect reconstruction filter banks are addressed, and two design approaches based on Jordan decomposition of the state transition matrix are presented. The analysis and synthesis polyphase matrices are constrained to be minimal. In the first approach, the state transition matrix of the synthesis polyphase is obtained by similarity transformation of the analysis state transition matrix. In the second approach, the state transition matrix of the synthesis polyphase is an analytic function of the analysis state transition matrix. Comparison of the two methods are provided with simulation results.

• P. Viswanadha Reddy, Anamitra Makur, Design of M Channel Bi Orthogonal FIR Perfect Reconstruction Filter Banks Using Spectral theory of Matrix Polynomials, International Conference on Information, Communications and Signal Processing, ICICS, IEEE - PCM, Singapore, Dec- 2003.

Abstract : In this work, the problem of design of M channel biorthogonal FIR perfect reconstruction filter banks (FIRPRFB) is addressed using spectral theory of matrix polynomials. The present characterization covers a broader class of filter banks. Inverse problem of construction of a regular matrix polynomial with inverse given it’s spectral data is used for such a characterization. Some insightful results on reconstruction delay and length of synthesis filters are  presented, both of these parameters can be varied in the present characterization. Simulation results are presented for 2,3,4 channel designs.
 
 
• P. Viswanadha Reddy, Anamitra Makur, Design of M Channel Causal Stable IIR Perfect Reconstruction Filter Banks Using  Similarity Transformation, IEEE International Conference on Image Processing, ICIP - 2003, Barcelona, Spain, Sep/2003.

Abstract : In this work, the problem of design of M channel IIR causal stable perfect reconstruction filter banks is addressed, and design approach based on similarity transformation of the state transition matrix is presented. The analysis and synthesis polyphase matrices are constrained to be minimal. The state transition matrix of the synthesis polyphase is obtained by  similarity transformation of the analysis state transition matrix. Simulation results are provided.
 
 
• P. Viswanadha Reddy, Anamitra Makur, M Channel Causal Stable IIR Perfect Reconstruction Filter Bank Design Using Minimal Factorization, IEEE International Conference on Acoustics Speech and Signal Processing, ICASSP - 2004, Montreal, Canada. submitted

Abstract : In this work, the problem of design of M channel IIR causal stable perfect reconstruction filter banks is addressed and a design approach based on the minimal factorization of the state space form of the polyphase matrix are presented based on the concept of factorization of rational matrix functions. The analysis and synthesis polyphase matrices are constrained to be minimal. Simulation results are presented for 6 channel and 2 channel cases. In this work, the problem of design of M channel IIR causal stable perfect reconstruction filter banks is addressed and a design approach based on the minimal factorization of the state space form of the polyphase matrix are presented based on the concept of factorization of rational matrix functions. The analysis and synthesis polyphase matrices are constrained to be minimal. Simulation results are presented for 3 channel and 4 channel cases.
 
 
• Narendra Asnani, Preeti Pisupati, P. Viswanadha Reddy, Sheri George, Data reduction as a technique for visualization of large data volume, NDE - 2003, Thiruvananthapuram, India, Dec - 2003.

Abstract : Nondestructive inspections of complex parts using advanced digital techniques such as computed tomography (CT) result in large data volumes. Image review time is a major bottleneck in analyzing such data. Commercially available packages to analyze and interpret large data sets on personal computer platforms are not very mature. These packages either do not provide the capability of handling large data volumes, or compromise bit-resolution in doing so. This paper discusses data reduction techniques, such as spatial averaging and sub-sampling, which do not compromise bit-resolution. The data reduction techniques are applied to industrial CT data (where several 2D slices are stacked to obtain a 3D model). Results are presented showing good rendering speeds for large data volumes. This paper also discusses the merits and limitations of the technique.