Design of Optimal Generalized Stack Filters under the Mean Absolute Error Criterion

Moncef Gabbouj,Bing Zeng,Yrjö Neuvo

1991-01-01

Abstract:Design of Optimal Generalized Stack Filters under the Mean Absolute Error Criterion.

What problem does this paper attempt to address?

Design of Minimum MAE Generalized Stack Filters for Image Processing

B ZENG,M GABBOUJ,Y NEUVO

DOI: https://doi.org/10.1117/12.50360

1991-01-01

Abstract:Design of optimal generalized stack filters (GSFs) under the mean absolute error (MAE) criterion suffers from two bottlenecks, that is, the design procedure depends on the joint statistics of the signal and noise processes that are rarely known and calls for a huge linear program (LP). In this paper we suggest efficient approaches to solve these problems. First, we present a method of estimate, based on training sequences, all the probabilities needed during the filter design procedure. Then, we introduce an algorithm that only involves data comparisons exclusively, but results in optimal filters in most practical cases. Design examples for image restoration from impulsive noise are provided.
Optimal Parallel Stack Filtering under the Mean Absolute Error Criterion.

Bing Zeng,Yrjö Neuvo

DOI: https://doi.org/10.1109/83.287029

1994-01-01

Abstract:The authors extend the configuration of stack filtering to develop a new class of stack-type filters called parallel stack filters (PSFs). As a basis for the parallel stack filtering, the block threshold decomposition (BTD) is introduced, and its properties are investigated. The design of optimal PSHs under the mean absolute error (MAE) criterion is shown to be similar to the minimum MAE stack fil...
Multilevel Stack Filtering for Image Processing Applications

T SUN,B ZENG,Y NEUVO

DOI: https://doi.org/10.1117/12.58366

1992-01-01

Abstract:In this paper, an optimal multilevel stack filtering algorithm (the stack filters used at each level are designed to be optimal) is introduced. Optimal multilevel stack filter is capable of considering a larger window mask than an optimal stack filter. Multilevel unidirectional stack filters and multilevel bidirectional stack filters are studied in particular. Design of an optimal multilevel stack filter is similar to the design of an optimal stack filter where the compare-and- select algorithm can be used. However, the design procedure needs to be adapted to operation on more than one level. Several design examples are presented showing the good performance of the proposed method.
Optimization of Stack Filters Based on Mirrored Threshold Decomposition

CUI Ying,ZHAO Chun-hui

DOI: https://doi.org/10.3969/j.issn.1006-7043.2006.06.025

2006-01-01

Abstract:Optimizing stack filters is the most pivotal issue of the study of stack filters theory.The threshold decomposition is one of stack filters property and has two modes: traditional threshold decomposition and mirrored threshold decomposition.The stack filters defined in the binary domain of mirrored threshold decomposition have been empowered not only with lowpass filtering characteristics but with bandpass and highpass characteristics as well.In this paper,under MAE criterion,a fast adaptive optimization algorithm based on mirrored threshold decomposition is presented.This algorithm restricts stacking properties of cost vector and estimates its astringency in order to attain positive Boolean function of optimal stack filters.The performance of optimal stack filters is illustrated by simulations.The results of experiments show that the optimal stack filters can suppress noise and protect the details of image effectively,which improve filtering capability.
Optimizing Stack Filters by Adaptive Whole Annealing Genetic Algorithm

ZHAO Chun-hui,CUI Ying,LI Gang

DOI: https://doi.org/10.3969/j.issn.1006-7043.2006.01.029

2006-01-01

Abstract:Stack filters are a class of nonlinear digital filters that satisfy a weak superposition property known as the threshold decomposition and stacking property.Their design is formulated as a course of optimizing positive Boolean function(PBF).Optimizing stack filters by an existing genetic algorithm easily relapses into a local best result.Due to this aspect,using an annealing mechanism as a selection operator is proposed.Cross probability and mutation probability were selected adaptively according to different individuals.Experiment results show that stack filters optimized by an adaptive whole annealing genetic algorithm(AWAGA) under mean squared error(MSE) criteria can suppress noise and protect the details of image effectively.
Design of Optimal Stack Filters: a Neural Net Approach with BP Algorithm

B Zeng

DOI: https://doi.org/10.1109/icsmc.1996.561377

1996-01-01

Abstract:It is well-known that the design of optimal stack filters has been restricted seriously by the filter's size. The maximum size that can now be handled with the developed techniques is 18 which cannot even cover fully a 5 x 5 square mask.In this paper, we present a neural network approach to the optimal design of stack filters where we treat each minimum (MIN) or maximum (MAX) operation as a neuron. In this way, we can design the positive Boolean function (PBF) directly, thus avoiding the determination of the whole Karnaugh map (which may have a prohibitively large size) as required in the conventional methods.The design of an optimal filter is accomplished by using the back-propagation (BP) algorithm. Some specific characteristics concerning the design process will be addressed in details, accompanied by a few design examples so as to justify the proposed method.
FIR Stack Hybrid Filters

Zhou, H.,Bing Zeng,Neuvo, T.

DOI: https://doi.org/10.1109/iscas.1991.176311

1991-01-01

Abstract:A recently developed approach to synthesizing optimal stack filters under the mean absolute error (MAE) criterion (see B. Zong et al., 1991) is extended to FIR (finite impulse response) stack hybrid (FSH) filters, where FIR filters are assumed to be the average filter. This approach is demonstrated through synthesizing a group of optimal FSH filters in the minimum MAE sense for the standard image `Bridge', where the task is to restore the image from impulsive noise. Practical testing results show that optimal FSH filters can do much better than the corresponding FIR median hybrid filters
Optimizing Stack Filters Through a Clone Selection Algorithm

ZHAO Chun-hui,SUN Li,FU Zheng-wei

DOI: https://doi.org/10.3969/j.issn.1006-7043.2007.04.018

2007-01-01

Abstract:Stack filters are a class of nonlinear filters that have good performance in suppressing noises and preserving details.They have recently become the hotspot of research works.Their design is actually an optimization problem of positive Boolean function(PBF).This paper presents a clone selection algorithm for optimizing stack filters under mean absolute error(MAE) criterion.Results of experiments show that stack filters optimized with a clone selection algorithm are more effective,and optimizing speed is better than that with filters optimized by genetic algorithms or by particle swarm algorithms,which indicates that the clone selection algorithms are effective for optimizing stack filters.
Adaptive Neighbor Weight Simulated Annealing Stack Filters

CUI Ying,ZHAO Chun-hui

DOI: https://doi.org/10.3969/j.issn.1006-9348.2010.11.060

2010-01-01

Abstract:The design of stack filters was formulated as a course of optimizing PBF,and how to effectively gain optimal Boolean function was difficult problem.In order to solve the above problem,simulated annealing algorithm(SA) was presented to design stack filters,whose optimization time was less than genetic algorithm,and the effect of neighbor pixel was considered,and the optimal mathematical model was constructed based on adaptive neighbor weight error(ANWMAE) criterion.The adaptive neighbor weight simulated annealing stack filters(ANWSA) and the adaptive neighbor weight recursive simulated annealing stack filters(RANWSA) were constructed based on the traditional threshold decomposition.In order to testify the filtering capability,ANWSA and RANWSA were simulated,the results show that they can suppress noise and protect the details of image effectively.
Steady-state performance analysis for adaptive filters with error nonlinearities

Bin Lin,Rongxi He, Liming Song,Baisuo Wang

DOI: https://doi.org/10.1109/ICASSP.2009.4960278

2009-01-01

Abstract:A unified approach to the steady-state mean square error (MSE) and tracking performance analyses for real and complex adaptive filtes with error nonlinearities is developed. Some general clofied-form analytical expressions for the steady-state performances are given. Our analyses are based on Taylor series expansion and and so-called complex Brandwood-form series expansion (BSE). Under these general explicit expressions, some well-known adaptive filters can be viewed as special cases. In addition, the closed-form analytical expressions for the steady-state performance for real and complex least-mean p-power (LMP) algorithm with different choices of parameter p are also given. A mass of simulations show the accuration of our analyses.
Optimal Stacked Filters by Simulated Annealing

CUI Ying,ZHAO Chun-hui

DOI: https://doi.org/10.3969/j.issn.1009-671x.2005.10.001

2005-01-01

Abstract:In order to improve optimization algorithm of staked filters,an algorithm based on simulated annealing(SA) is proposed to deal with grey images.Stacked filters are a class of nonlinear digital filters based on threshold decomposition,whose optimizing process is a course of optimizing PBF.SA is a conventional optimization algorithm,and is widely used in engineering.The consumed time is less than genetic algorithm.In this paper,SA is employed to optimize stacked filters.Experimental results indicate that the optimized stacked filters can eliminate noise and protect the details of image effectively.
Optimization Design of Soft Morphological Filters Based on Improving Genetic Algorithm

CH Zhao

DOI: https://doi.org/10.1109/icnnsp.2003.1279318

2002-01-01

Abstract:Soft morphological filters form a large subclass of stack filters. They were introduced to improve the behavior of standard morphological filters in noisy conditions. But design methods existing for these filters tend to be computationally intractable or require some special knowledge. Genetic algorithms provide useful tools for optimization problems. In this paper, a method of soft morphological filter design using improving genetic algorithm is described, and the behavior of soft morphological filters is illustrated by empirical results. Some optimal parameters are also proposed for MAE and MSE error criteria.
Far-infrared Stack Hybrid Filters

B ZENG,HB ZHOU,Y NEUVO

DOI: https://doi.org/10.1117/12.55889

IF: 1.3

1991-01-01

Optical Engineering

Abstract:We introduce a new class of linear-nonlinear combinational filters, called FIR stack hybrid (FSH) filters, which greatly enlarges the family of FIR median hybrid (FMH) filters. The optimal FSH filtering theory under the mean absolute error (MAE) criterion is studied. A general two-step method to synthesize optimal FSH filters is developed. In the first step, the probabilities needed in the optimal filter design are estimated based on images. In the second step, the linear program (LP) required to solve the best filter is avoided by using a good suboptimal algorithm that only needs data comparisons. A sufficient condition under which the suboptimal routine can result in optimal solutions is given, and this condition is shown to hold in most practical cases. To demonstrate the efficiency of the proposed approach, a group of FSH filters are synthesized for the standard image "Bridge." The task is to restore the image from impulsive or Gaussian noise. By checking the sufficient condition, each filter is found to be optimal in the minimum MAE sense. Testing results of some synthesized filters show that the optimal FSH filters can do a much better job than the corresponding FMH filters in both noise environments.
Optimal Design of Stack Filters Using an Improved Clonal Selection Algorithm

赵春晖,赵华,万建

DOI: https://doi.org/10.3969/j.issn.1673-4785.2008.03.012

2008-01-01

Abstract:The optimal design of stack filters is actually an optimization of a positive Boolean function.In order to speed up the optimizing rate and produce a global optimal design of the stack filters,an improved clonal selection algorithm(ICSA) is presented,which introduces a polyclonal operator and simultaneous evolution of the memory unit and the reserved group.The number of clonal particles in the memory unit can adaptively change with their affinity concentration.Regrouping operations are made between parent generations and child generations in the memory unit,avoiding problems caused by breeding two close particles.The preservation of group mutations guarantees diversity.Our experimental results confirmed that stack filters optimized with ICSA produce better filtering results in less time.
Stack Filters Based on Whole Annealing Fuzzy Genetic Algorithm

CUI Ying,ZHAO Chun-hui,TANG Chun-ming

DOI: https://doi.org/10.3969/j.issn.1673-9728.2006.02.296

2006-01-01

Abstract:In this paper,fuzzy control is applied in genetic algorithm to control crossover and mutation probability,at the same time,it introduces annealing mechanism to selection operator,we apply the whole annealing fuzzy genetic algo- rithm to optimize stack filters and analyze its filtering performance.The performance of optimal stack filters is iUustra- ted by simulations,results of experiments show that the stack filters designed by this algorithm can suppress noise and protect the details of image effectively,It improve filtering capability.
Stacking Learning with Coalesced Cost Filtering for Accurate Stereo Matching

Peng Yao,Jieqing Feng

DOI: https://doi.org/10.1016/j.jvcir.2021.103169

IF: 2.887

2021-01-01

Journal of Visual Communication and Image Representation

Abstract:Deep learning based stereo matching algorithms have produced impressive disparity estimation for recent years; and the success of them has once overshadowed the conventional ones. In this paper, we intend to reverse this inferiority, by leveraging Stacking Learning with Coalesced Cost Filtering to make the conventional algorithms achieve or even surpass the results of deep learning ones. Four classical and Discriminative Dictionary Learning (DDL) algorithms are adopted as base-models for Stacking. For the former ones, four classical stereo matching algorithms are employed and regarded as ‘Coalesced Cost Filtering Module’; for the latter supervised learning one, we utilize the Discriminative Dictionary Learning (DDL) stereo matching algorithm. Then three categories of features are extracted from the predictions of base-models to train the meta-model. For the meta-model (final classifier) of Stacking, the Random Forest (RF) classifier is selected. In addition, we also employ an advanced one-view disparity refinement strategy to compute the final refined results more efficiently. Performance evaluations on Middlebury v.2 and v.3 stereo data sets demonstrate that the proposed algorithm outperforms other four most challenging stereo matching algorithms. Besides, the submitted online results even show better results than deep learning ones.
Optimizing Stacked Filters with Mirrored Threshold Decomposition Using an Improved Clone Selection and PSO Algorithm

ZHAO Chun-hui,XIAO Xiao-jun

DOI: https://doi.org/10.3969/j.issn.1673-4785.2008.02.005

2008-01-01

Abstract:Stack filters are a class of nonlinear digital filters with sliding windows.Their two main properties are threshold decomposition and stacking.Optimal design of these filters requires optimization of a positive Boolean function(PBF).Optimization of stacked filters using traditional algorithms is easily trapped in a local optimum.To prevent this in recent years the bionic clone selection and PSO algorithms were proposed.The combination of these two algorithms avoids the problem of falling into a local minimum,greatly improving convergence speed and accuracy.Experimental results show that optimally stacked filters effectively suppress noise while retaining details of images and enhancing filtering capability.
Adaptive Weight Mirrored Threshold Stack Filters

崔颖,赵春晖,汤春明,张健沛

2009-01-01

Abstract:Compared to the traditional threshold stack filters,mirrored threshold stack filters have been empowered not only with lowpass filtering characteristics but with bandpass and highpass characteristics as well,but their positive Boolean functions length leads to an increasing restriction during calculation.In order to solve the above problems,this paper proposes a mirrored adaptive weight(MAW)algorithm,which calculates cost vector based on adaptive neighbor weight error criterion(ANWMAE).After cost vector stacking is restricted and its astringency is estimated,the optimal positive Boolean function of stack filters is confirmed to construct adaptive weight mirrored threshold stack filters(AWMSF).In order to testify the filtering capability,AWMSF was simulated,the results show that it can suppress noise and protect the details of image effectively,the number of iterations is reduced and the computing complication is decreased rapidly too.
Speckle reduction with adaptive stack filters

María Elena Buemi,Alejandro C. Frery,Heitor S. Ramos

DOI: https://doi.org/10.1016/j.patrec.2013.06.005

IF: 4.757

2014-01-01

Pattern Recognition Letters

Abstract:Stack filters are a special case of non-linear filters. They have a good performance for filtering images with different types of noise while preserving edges and details. A stack filter decomposes an input image into stacks of binary images according to a set of thresholds. Each binary image is then filtered by a Boolean function, which characterizes the filter. Adaptive stack filters can be computed by training using a prototype (ideal) image and its corrupted version, leading to optimized filters with respect to a loss function. In this work we propose the use of training with selected samples for the estimation of the optimal Boolean function. We study the performance of adaptive stack filters when they are applied to speckled imagery, in particular to Synthetic Aperture Radar (SAR) images. This is done by evaluating the quality of the filtered images through the use of suitable image quality indexes and by measuring the classification accuracy of the resulting images. We used SAR images as input, since they are affected by speckle noise that makes classification a difficult task.

computer science, artificial intelligence
Computationally efficient robust adaptive filtering algorithm based on improved minimum error entropy criterion with fiducial points

Xinyan Hou,Haiquan Zhao,Xiaoqiang Long,Hing Cheung So

DOI: https://doi.org/10.1016/j.isatra.2024.04.008

IF: 7.3

2024-04-11

ISA Transactions

Abstract:Highlights • In this paper, we propose an improved minimum error entropy criterion with fiducial points for the high computational complexity of the minimum error entropy criterion with fiducial points (MEEF), and derive a gradient-based adaptive filtering algorithm based on the improved criterion. • We analyze the convergence conditions, transient and steady state behaviors of the proposed algorithm based on statistical properties. • The experimental results show that the proposed algorithm ensures the accuracy of the algorithm while greatly reducing the computational complexity of the algorithm and the correctness of the theoretical model. Recently, there has been a strong interest in the minimum error entropy (MEE) criterion derived from information theoretic learning, which is effective in dealing with the multimodal non-Gaussian noise case. However, the kernel function is shift invariant resulting in the MEE criterion being insensitive to the error location. An existing solution is to combine the maximum correntropy (MC) with MEE criteria, leading to the MEE criterion with fiducial points (MEEF). Nevertheless, the algorithms based on the MEEF criterion usually require higher computational complexity. To remedy this problem, an improved MEEF (IMEEF) criterion is devised, aiming to avoid repetitive calculations of the aposteriori error, and an adaptive filtering algorithm based on gradient descent (GD) method is proposed, namely, GD-based IMEEF (IMEEF-GD) algorithm. In addition, we provide the convergence condition in terms of mean sense, along with an analysis of the steady-state and transient behaviors of IMEEF-GD in the mean-square sense. Its computational complexity is also analyzed. Simulation results demonstrate that the computational requirement of our algorithm does not vary significantly with the error sample number and the derived theoretical model is highly consistent with the learning curve. Ultimately, we employ the IMEEF-GD algorithm in tasks such as system identification, wind signal magnitude prediction, temperature prediction, and acoustic echo cancellation (AEC) to validate the effectiveness of the IMEEF-GD algorithm.

automation & control systems,instruments & instrumentation,engineering, multidisciplinary

Design of Optimal Generalized Stack Filters under the Mean Absolute Error Criterion

Design of Minimum MAE Generalized Stack Filters for Image Processing

Optimal Parallel Stack Filtering under the Mean Absolute Error Criterion.

Multilevel Stack Filtering for Image Processing Applications

Optimization of Stack Filters Based on Mirrored Threshold Decomposition

Optimizing Stack Filters by Adaptive Whole Annealing Genetic Algorithm

Design of Optimal Stack Filters: a Neural Net Approach with BP Algorithm

FIR Stack Hybrid Filters

Optimizing Stack Filters Through a Clone Selection Algorithm

Adaptive Neighbor Weight Simulated Annealing Stack Filters

Steady-state performance analysis for adaptive filters with error nonlinearities

Optimal Stacked Filters by Simulated Annealing

Optimization Design of Soft Morphological Filters Based on Improving Genetic Algorithm

Far-infrared Stack Hybrid Filters

Optimal Design of Stack Filters Using an Improved Clonal Selection Algorithm

Stack Filters Based on Whole Annealing Fuzzy Genetic Algorithm

Stacking Learning with Coalesced Cost Filtering for Accurate Stereo Matching

Optimizing Stacked Filters with Mirrored Threshold Decomposition Using an Improved Clone Selection and PSO Algorithm

Adaptive Weight Mirrored Threshold Stack Filters

Speckle reduction with adaptive stack filters

Computationally efficient robust adaptive filtering algorithm based on improved minimum error entropy criterion with fiducial points