Mitsuji Muneyasu

Synthesis of 2-D separable-denominator digital filters with low sensitivity

Abstract This paper treats the coefficient sensitivity of two-dimensional (2-D) separable-denominator digital filters using the Roesser local state-space model. Two techniques suitable for 2-D separable-denominator digital filters are developed for synthesizing the filter structure with low sensitivity, one free from 12 scaling constraints on the state variables, and the other under the scaling constraints. In the paper, it is clarified that the filter structures with low sensitivity can be easily derived from the balanced realization. Finally, an example is given to illustrate the utility of the proposed techniques.

Edge-preserving smoothing by adaptive nonlinear filters based on fuzzy control laws

This paper proposes a novel type of edge-preserving smoothing filters to he applied to the images corrupted with impulsive and white Gaussian noise. This filter is based on the weighted mean filter whose coefficients can be varied adaptively by some kinds of local features in the window. The fuzzy control laws can be used for the implementation of the proposed filter and the optimization technique for the parameters of the membership functions and the fuzzy rules are described. Finally, the simulation example is given to illustrate the effectiveness of the proposed technique.

A new adaptive center weighted median filter using counter propagation networks

Abstract A new type of adaptive center weighted median filters is developed for impulsive noise reduction of an image without the degradation of an original signal. The weight in the proposed filter is decided by the weight controller based on counter propagation networks. This controller classifies an input vector into some cluster according to its feature and gives the weight corresponding to the cluster. The parameters in the weight controller are adjustable by using the learning algorithm. The degradation of the original signal can be reduced by the proposed technique. An example is also given to illustrate the utility of the proposed technique.

A novel 2-D adaptive filter based on the 1-D RLS algorithm

This paper proposes a novel two-dimensional (2-D) adaptive filter by applying a 1-D recursive least-squares (RLS) algorithm along both horizontal and vertical directions. The relation of the proposed algorithm to a usual 2-D RLS algorithm are investigated. A method that employs a priori estimation error is also considered to accelerate the convergent rate of the algorithm. The proposed filter has a good performance in nonstationary case, and the accuracy of convergence is better than in the existing 2-D least mean square (LMS) adaptive filters. The amount of computations required for the proposed algorithm are relatively small. Finally, an example is given to illustrate the utility of the proposed filter.

A realization of 2-D adaptive filters using affine projection algorithm

Abstract This paper proposes a realization of two-dimensional (2-D) adaptive finite impulse response (FIR) filters using affine projection method. The convergence property of the proposed algorithm is superior to that of the 2-D least mean square (LMS) and normalized LMS (NLMS) algorithms. This algorithm has properties that lie between those of NLMS and recursive least squares (RLS) algorithms. The convergence of this filter can be proved. The generalization of the proposed algorithm is also discussed. To illustrate the utility of the proposed technique, this filter is applied to the 2-D system identification.

2-D adaptive state-space filters based on the Fornasini-Marchesini second model

Based on the Fornasini-Marchesini second model, a technique is developed for implementing two-dimensional (2-D) adaptive state-space filters. First, the relationship between the coefficient sensitivities and the intermediate transfer functions is investigated for the Fornasini-Marchesini second model. A least mean square (LMS) adaptive algorithm is then presented by using new systems that generate the gradient signals. Finally, a 2-D adaptive line enhancer is constructed by using the 2-D adaptive state-space filter to illustrate the utility of the proposed technique.

A pipeline architecture of quadratic adaptive Volterra filters based on NLMS algorithm

This paper proposes a pipeline implementation of quadratic adaptive Volterra filters based on NLMS algorithm for high-speed processing and low power consumption. This implementation also gives a superior convergence performance than that of LMS pipeline algorithm under the colored inputs condition. The proposed implementation has zero output latency and reduced critical path compared to the non-pipelined implementation.

2-D adaptive state-space digital filters using LMS algorithm

This paper treats the realization of 2-D adaptive state-space filters using the least mean squares (LMS) algorithm. This is based on the Roesser local state-space model. First, the relation between the coefficient sensitivities and the intermediate transfer functions is investigated. Then, after introducing new systems for generating the gradient vectors, an adaptive algorithm is developed using the LMS algorithm. Finally, in a numerical example the 2-D adaptive state-space filter is used to design a 2-D digital filter in the spatial domain.<<ETX>>

A new 2-D adaptive filter using affine projection algorithm

This paper proposes a realization of two-dimensional (2-D) adaptive finite impulse response (FIR) filters using the affine projection method. The convergence property of the proposed algorithm is superior to that of the 2-D least mean square (LMS) and normalized LMS (NLMS) algorithms. This algorithm has properties that lie between those of NLMS and recursive least squares (RLS) algorithms. The generalization of the proposed algorithm is also discussed. To illustrate the utility of the proposed technique, this filter is applied to the 2-D system identification.

Improvement of the detection method for carotid artery calcification in dental panoramic radiographs

This paper proposes improvement for the detection method for carotid artery calcification in dental panoramic radiographs. The existence of the carotid artery calcification as an index for arteriosclerosis attracts a great deal of attention. The detection method for carotid artery calcification has been proposed. However the detection rate is about 50% for real data and its improvement should be required. In this paper, the area of the processing window is narrowed to be 300 × 300 pixels. Then we focus on the shape and area of the target region and improve the detection and misdetection rates. From experimental results, the carotid artery calcification region can extract more precisely compared to the conventional method.