Image Processing with Order Statistic and Morphological Filters

Chair: Edward J. Coyle, Purdue University, USA

Home


Conditional Morphology for Image Restoration

Authors:

Carlo S. Regazzoni, University of Genoa (Italy)
E. Stringa, University of Genoa (Italy)
C. Valpreda, University of Genoa (Italy)

Volume 1, Page (NA), Paper number 441

Abstract:

In this paper some properties of conditional morphology for image restoration are presented. Conditional morphology is a nonlinear method for coupled edge detection and image smoothing and it is derived from the fusion of two nonlinear Bayesian approaches: statistical morphology and the deterministic annealing solution to Markov Random Fields models. The proposed method resulting from fusing these approaches, is based on Mean Field Theory, which represents a common theoretical framework able to provide a computational simplification to the problem solution together with maintaining the robustness of Bayesian models. Experimental results are presented showing good performances obtained by the proposed approach on image affected by impulsive multiplicative noise (e.g. speckle noise).

ns970441.pdf (Scanned)

ns970441.pdf (From Postscript)

TOP



On the Design of Adaptive Fuzzy Regression Selection Filter with the Application of Image Restoration

Authors:

Pao-Ta Yu, National Chung Cheng University (Taiwan)
Jia-Hong Wu, National Chung Cheng University (Taiwan)

Volume 1, Page (NA), Paper number 442

Abstract:

In this paper we present a detection-estimation filter based on the technique of adaptive fuzzy regression. The method of fuzzy regression is useful in data estimation and is more flexible than the conventional regression methods. However, the time to calculate an optimal solution in fuzzy regression analysis prohibits us to apply the fuzzy regression method directly to the field of image restoration. Thus, we present a new algorithm to approximate the optimal solution of the fuzzy linear regression method. This new method, called the adaptive fuzzy linear regression method, will meet the need of speed in image restoration. The Adaptive Fuzzy Regression Selection (AFRS) filter combines the technique of the adaptive fuzzy linear regression with the RCRS filter. The central pixel of the window in the AFRS filter is replaced with the estimated output of the RCRS filter only if the central pixel is corrupted. A model is formed using the adaptive fuzzy linear regression method to decide whether the central pixel is corrupted. The computer simulations prove that the A FRS filter is useful in image restoration.

ns970442.pdf (Scanned)

ns970442.pdf (From Postscript)

TOP



Perceptually Optimal Restoration of Images with Stack Filters

Authors:

Jr-Jen Huang, Purdue University (U.S.A.)
Edward J. Coyle, Purdue University (U.S.A.)

Volume 1, Page (NA), Paper number 443

Abstract:

The present approach to the MAE-based design of stack filters for image restoration does not always produce the desired visual result. Thus, in this paper, a new stack filter design algorithm is developed. It is based upon a Weighted Mean Absolute Error (WMAE) criterion instead of the traditional MAE criterion, which assigns the same weights to all errors. The weights in this WMAE criterion are designed with. the aid of the Visible Differences Predictor (VDP), which can estimate the sensitivity of the human visual system to changes in images. Experiments with this WMAE approach show that the stack filters it produces perform significantly better in image processing applications than those designed with the MAE approach.

ns970443.pdf (Scanned)

ns970443.pdf (From Postscript)

TOP



Nonlinear Filter Design Based on Fuzzy Inference Rules for Image Processing

Authors:

Rong-Chung Cheng, National Chung Cheng University (Taiwan)
Pao-Ta Yu, National Chung Cheng University (Taiwan)

Volume 1, Page (NA), Paper number 444

Abstract:

A fuzzy filtering system is proposed for image restoration. Since all the conventional filters have their specific characteristics, they act well for some environments but with poor performance for others. The fuzzy filtering system gives the method to aggregate the advantage of the conventional filters to obtain the improved performance for image restoration.

ns970444.pdf (Scanned)

ns970444.pdf (From Postscript)

TOP



Image Sharpening by Morphological Filtering

Authors:

J.G.M. Schavemaker, Delft University of Technology (The Netherlands)
M.J.T. Reinders, Delft University of Technology (The Netherlands)
R. van den Boomgaard., University of Amsterdam (The Netherlands)

Volume 1, Page (NA), Paper number 445

Abstract:

In this paper we introduce a class of morphological operators with applications to sharpening digitized grey valued images. We introduce the underlying partial differential equation (PDE) that governs this class of operators. For discrete implementations of the operator class, we show that instances utilizing a parabolic structuring function, have special properties that lead to an efficient implementation and isotropic sharpening behavior.

ns970445.pdf (Scanned)

ns970445.pdf (From Postscript)

TOP



A Comparative Study of the Complexity of Multivariate Medians

Authors:

Mauro Barni, University of Florence (Italy)
Vito Cappellini, University of Florence (Italy)

Volume 1, Page (NA), Paper number 446

Abstract:

Multivariate median filters represent a powerful tool for edge preserving noise removal from multichannel digital images. However, the usability of such filters in practical applications is often limited because of their high computational complexity, all the more that a comprehensive analysis of the complexity of the various classes of multivariate medians is still missing. In this work, the complexity of many multivariate extensions of the median filter is briefly discussed. Both theoretical analysis and experimental results show that the computational complexity depends mainly on the strategy adopted to sort multivariate data. The use of marginal ordering leads to the fastest algorithms, filters relying on reduced ordering have an intermediate behavior, whereas those based on aggregate ordering are by far the most complex.

ns970446.pdf (Scanned)

ns970446.pdf (From Postscript)

TOP