Novel Convolutions Using First-Order Moments

This paper presents a novel fast algorithm for digital convolutions. It is able to compute arbitrary-length convolutions more efficiently via transforming the convolution into a first-order moment. Although many additions are required, the proposed algorithm has some advantages such as the avoidance of multiplications, simple computation structure, and only integer additions. These advantages contribute to this algorithm being so easy that it can compute convolutions rapidly. Based on the proposed algorithm a very simple and scalable systolic array without multipliers and ROM has been developed leading to more efficient VLSI implementation of convolutions.

[1]  Vladik Kreinovich,et al.  Fast convolution and Fast Fourier Transform under interval and fuzzy uncertainty , 2010, J. Comput. Syst. Sci..

[2]  Pramod Kumar Meher,et al.  Hardware-Efficient Systolization of DA-Based Calculation of Finite Digital Convolution , 2006, IEEE Transactions on Circuits and Systems II: Express Briefs.

[3]  H. Nussbaumer,et al.  Fast polynomial transform algorithms for digital convolution , 1980 .

[4]  David M. Mount,et al.  Approximating large convolutions in digital images , 1998, Optics & Photonics.

[5]  P. Thangaraj,et al.  Low Power Shift and Add Multiplier Design , 2010, ArXiv.

[6]  Chein-Wei Jen,et al.  A memory efficient realization of cyclic convolution and its application to discrete cosine transform , 2003, Proceedings of the 2003 International Symposium on Circuits and Systems, 2003. ISCAS '03..

[7]  F. K. Lam,et al.  A Novel Approach to Fast Discrete Fourier Transform , 1998, J. Parallel Distributed Comput..

[8]  Richard E. Blahut,et al.  Fast Algorithms for Digital Signal Processing , 1985 .

[9]  Chein-Wei Jen,et al.  Distributed arithmetic realisation of cyclic convolution and its DFT application , 2005 .

[10]  Pramod Kumar Meher,et al.  Parallel and Pipelined Architectures for Cyclic Convolution by Block Circulant Formulation Using Low-Complexity Short-Length Algorithms , 2008, IEEE Transactions on Circuits and Systems for Video Technology.

[11]  Madihally J. Narasimha.  Linear Convolution Using Skew-Cyclic Convolutions , 2007, IEEE Signal Processing Letters.

[12]  T. Lundy,et al.  A new matrix approach to real FFTs and convolutions of length 2k , 2007, Computing.

[13]  J. Cooley,et al.  New algorithms for digital convolution , 1977 .

[14]  K.K. Parhi,et al.  Hardware Efficient Fast DCT Based on Novel Cyclic Convolution Structures , 2006, IEEE Transactions on Signal Processing.

[15]  Madihally J. Narasimha.  Modified Overlap-Add and Overlap-Save Convolution Algorithms for Real Signals , 2006, IEEE Signal Processing Letters.

[16]  Okan K. Ersoy.  Semisystolic Array Implementation of Circular, Skew Circular, and Linear Convolutions , 1985, IEEE Transactions on Computers.

[17]  Thanos Stouraitis,et al.  Systolic algorithms and a memory-based design approach for a unified architecture for the computation of DCT/DST/IDCT/IDST , 2005, IEEE Transactions on Circuits and Systems I: Regular Papers.

[18]  F. H. Y. Chan,et al.  An all adder systolic structure for fast computation of moments , 1996, J. VLSI Signal Process..