Mitra Sanjit - Selected Publications#
1. “A new approach to the realization of low sensitivity IIR digital filters,” IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 34, 1986, pp. 350-361 (with P.P. Vaidyanathan and Y. Neuvo)
– Advanced the theory for realizing infinite-impulse response (IIR) digital filters with low sensitivity to changes in filter coefficients in the passband and proposed a method to design such filters with the fewest number of digital multipliers which are the most important part of the filter consuming large amount of powers and occupying the largest amount of space on a chip. The design method has been included in the Signal Processing Toolbox of MATLAB and the LabView Software package of National Instruments Inc.
2. “Interpolated finite impulse response digital filters,” IEEE Trans. on Acoustics, Speech, and Signal Processing, vol. 32, 1984, pp. 563-570 (with Y. Neuvo and C-Y Dong)
– Advanced a method of designing finite-impulse response (FIR) digital filters with the fewest number of multipliers. The design method has been included in the Signal Processing Toolbox of MATLAB and the LabView Software package of National Instruments Inc.
3. “A new efficient approach for the removal of impulse noise from highly corrupted images,” IEEE Trans. on Image Processing, Special Issue on Nonlinear Image Processing, vol. 5, 1996, pp. 1012-1025 (with E. Abreu, M. Lightstone and K. Arakawa)
- In images captured by a digital camera or in the transmission of an image over a communication channel, a number of pixels get corrupted by impulse like noises. Existing methods based on median filtering and order-statistics assume the pixels to be independent. A new framework that takes into account the correlation among neighboring pixels for removing impulse noise from images is advanced in which the nature of the filtering operation is conditioned on a state variable defined as the output of a classifier that operates on the differences between the input pixel and the remaining rank-ordered pixels in a sliding window. As part of this framework, several algorithms have been examined. First, a simple two-state approach is described in which the algorithm switches between the output of an identity filter and a rank-ordered mean (ROM) filter. For a small additional cost in memory, this simple strategy has been easily generalized into a multistate approach using weighted combinations of the identity and ROM filter in which the weighting coefficients are optimized using image training data. Extensive simulations have shown that the proposed methods perform significantly better than a number of existing nonlinear techniques with as much as 40% impulse noise corruption. Moreover, the method can effectively restore images corrupted with Gaussian noise and mixed Gaussian and impulse noise.
4. “Block implementation of adaptive digital filters,” IEEE Trans. Acoustics, Speech, Signal Proc., vol. 29, 1981, pp. 744-752 (with G. Clark and S. Parker)-
- Block digital filtering involves the calculation of a block or finite set of filter outputs from a block of input values. This type of filtering leads to a very general family of filters that includes both the conventional scalar implementation and batch processing, in addition to providing a reduction in sensitivity to roundoff and coefficient accuracy. This paper presents a block adaptive filtering procedure in which the filter coefficients are adjusted once per each output block in accordance with a generalized least mean-square (LMS) algorithm. Analyses of convergence properties and computational complexity show that the block adaptive filter permits fast implementations while maintaining performance equivalent to that of the widely used LMS adaptive filter.
5. “Tunable digital frequency response equalization filters,” IEEE Trans. on Acoustics, Speech, and Signal Processing, vol. 35, 1987, pp. 118-120 (with P.A. Regalia)-
- A tunable second-order digital filter structure with parametrically tunable magnitude response is advanced. For this structure, each of the following parameters, the gain of the bandpass response or the loss of the band-stop response, the frequency at which the peak or the dip of the magnitude response, and the 3-dB bandwidth of the magnitude response can be changed by an individual multiplier coefficient. This type of filter is an important component in a digital audio workstation with a mixing board for producing and recording music, songs, sound effects and other situations where recorded audio is needed.
6. “Nonuniform discrete Fourier transform and its applications in filter design, Parts I and II,” IEEE Trans. on Circuits & Systems: Part II, vol. 43, 1996, pp. 422-433 and pp. 434-444 (with S. Bagchi)
Developed the frequency representation of a finite-length digital sequence at arbitrarily spaced points in the complex-frequency domain. This type of representation has many practical applications including antenna array design, detection of dual-tone multi-frequency (DTMF) tones, magnetic resonance imaging, numerical solution of partial differential equations and spectral analysis.
7. “Structural subband decomposition of sequences and its signal processing applications,” IEE Proceedings - Vision, Image & Signal Processing, vol. 146, 1999, pp. 109-123 (with U. Heute).
– Proposed a modified version of the well-known discrete Fourier transform (DFT) of a finite-length sequence by decomposing the sequence into subsequences with spectral separation property allowing a faster computation of the DFT of the dominant subsequence. This paper was awarded the Blumlein-Browne-Willans Premium (2000) of IEE (London).
8. “Warped discrete-Fourier transform: Theory and applications,” IEEE Trans. on Circuits & Systems, Part I, vol. 48, 2001, pp. 1086-1093 (with A. Makur)
– Advanced another modified version of the well-known discrete Fourier transform (DFT) for the development of the frequency domain representation of unequally spaced digital samples of a sequence in the complex-frequency domain. This type of DFT has several practical applications such as computationally efficient spectral analysis, design of tunable finite-impulse response digital filters, detection of dominant sinusoidal signal from a signal composed of closely spaced sinusoids, and design of perfect reconstruction filter banks with nonuniformly spaced passbands of filters in the bank, frequency estimation of noise-corrupted closely-spaced sinusoid and perceptual speech enhancement.
9. “A unified rate-distortion analysis framework for transform coding,” IEEE Trans. on Circuits & Systems for Video Technology, vol.11, 2001, pp. 1221-1236 (with Z. He)-
A unified rate-distortion (R-D) modeling framework for H.263 video coding based on the new concepts of characteristic rate (R) curves and rate curve decomposition (D) was advanced. Based on this framework, a unified R-D estimation and control algorithm is advanced for all typical image/video transform coding systems, such as embedded zero-tree wavelet (EZW), set partitioning in hierarchical trees (SPIHT) and JPEG image coding; MPEG-2, H.263, and MPEG-4 video coding. A linear rate regulation scheme has also been advanced to further improve the estimation accuracy and robustness, as well as to reduce the computational complexity of the R-D estimation algorithm. Extensive experimental results have verified the proposed algorithm. Received the 2001 IEEE Transactions on Circuits and Systems for Video Technology Best Paper Award.
10. “Noncausal filters in multipath channel shaping,” (with E. Abreu and R. Marchesani), U.S. Patent No. 5,533,063, July 2, 1996
– Advanced a practical method for shaping the channel impulse response (CIR) of a non-minimum phase communications channel, by reversing the CIR without enhancing the noise level using non-causal allpass filters is disclosed. Unstable allpass filters are implemented as stable non-causal filters operating in reversed time to obtain a minimum phase response from a non-minimum phase CIR. The original CIR is estimated using adaptive algorithms, and the coefficients of the allpass filters are taken directly from the estimated CIR. The cascade of the channel and allpass filter has an impulse response that is the approximate time reversed version of the original CIR. A “block” allpass equalizer and an “optimal” allpass equalizer are used in a two-stage filter to increase system performance. Other multiple stage filter configurations are disclosed, which can include use of decision feedback equalizer (DFE) decoders as well as Viterbi decoders. Licensed by Telettra/Alcatel Corporation and has been included in one of their equipment.
