Search

Kaniadakis entropy is a kind of generalized entropy based on the $$ \kappa $$ probability distribution, which has a good ability to deal with the distribution of long tail. The image thresholding algorithm based on Kaniadakis entropy can effectively segment images with long-tailed distribution histograms, such as nondestructive testing image. However, Kaniadakis entropy is a generalized information entropy with parameter. How to choose appropriate parameter $$ \kappa $$ is a problem to be solved. In this paper, we proposed an adaptive parameter selection Kaniadakis entropy thresholding algorithm. Based on a clustering effectiveness evaluation index, we transform the parameter selection problem into an optimization problem, then use particle swarm optimization search algorithm to optimize it and finally obtain the segmentation threshold under the optimal parameter. The presented algorithm can adaptively select parameters according to different images and obtain the optimal segmentation images. In order to show the effectiveness of the proposed method, the segmentation results are compared with several existing entropy-based thresholding algorithms. Experimental results both qualitatively and quantitatively demonstrate that the proposed method is effective.

In order to reduce the computational complexity of multi-objective evolutionary optimization-based clustering algorithms, a Kriging-assisted reference vector guided multi-objective robust spatial fuzzy clustering algorithm (KRV-MRSFC) is proposed and then successfully applied to image segmentation. We first construct objective functions with noise robust local spatial information derived from the image to improve the robustness to noise and then use the Kriging model to approximate each objective function to decrease the computational cost. Meanwhile, in order to improve the approximation accuracy of the Kriging model, an angle-penalized distance-based expected improvement sampling criterion is presented in the KRV-MRSFC, which can select individuals with better exploitation and exploration to update the Kriging model. In addition, KRV-MRSFC adopts a clustering validity index with noise robust local image spatial information to select the optimal solution from the final non-dominated solution set to perform image segmentation. The experiments performed on Berkeley and real magnetic resonance images indicate that the proposed method not only achieves satisfactory segmentation performance on noisy images but also requires a low time cost.

Crime scene investigation (CSI) image retrieval plays an important role in police case solving. In order to relieve the problem of over-fitting in the training of Convolutional Neural Network (CNN) model due to the limited number of CSI images, this paper presents a CNN feature fusion method based on transfer learning algorithm. Firstly, the pre-trained models VGG-F and VGG-Very Deep Convolutional Network 16 (VGG-VD 16) are fine-tuned, then the features of Fully Connected Layer 7 (FC7) are extracted from the two models respectively, and Principal Component Analysis (PCA) method is used to reduce the features dimensions. Then the features from both models are fused as the final image feature. Tested on CSI image dataset, experimental results show that the proposed algorithm has very good performance.

Fuzzy partition entropy-based method is an effective way for image segmentation. In this paper a segmentation method based on a weak fuzzy partition is presented. Firstly, we propose a method to construct generalized fuzzy complement, and moreover construct a generalized fuzzy complement operator which has a nice property for parameter optimization in real application. Then, a one-dimensional (1D) weak fuzzy partition, a two-dimensional (2D) weak fuzzy partition being obtained by a Cartesian product of two 1D fuzzy partitions, are defined using the proposed generalized fuzzy complement. With these concepts, a weak fuzzy partition entropy-based image segmentation method is proposed. The method is described in the 1D and 2D cases by modeling the 1D and 2D histograms. The 2D approach allows us to ensure a spatial regularity of the fuzzy classification. Finally, a nested optimization method is developed, based on an improved uniformity measure, to search for the optimal threshold in the image segmentation method. Empirical results show that the proposed weak fuzzy partition entropy-based method is capable of achieving better segmentation results than several state-of-the-art methods that are based on or not based on fuzzy entropy. The proposed 2D weak fuzzy partition entropy-based method is especially effective for noisy images.

In view of local feature weighting hard c-means (LWHCM) clustering algorithm sensitive to noise, based on a non-Euclidean metric, a robust local feature weighting hard c-means (RLWHCM) clustering algorithm is presented. RLWHCM is a natural, effective extension of LWHCM. The robustness of RLWHCM is analyzed by using the location M-estimate in robust statistical theory. The convergence proof of RLWHCM is given. Experimental results on synthetic and real-world data sets demonstrate the effectiveness of the proposed algorithm.

Image thresholding based on rough entropy is an efficient image segmentation technique. The optimal thresholds of the existing exponential and logarithmic rough entropy thresholding segmentation algorithms have unclear physical meaning. In this paper, a new form of square rough entropy is defined to measure the roughness in an image, and the corresponding image thresholding segmentation algorithm is proposed. The novel square rough entropy has good properties and simple computation. In the proposed thresholding algorithm, the optimal threshold is at the boundary between the object and the background of an image. This process is consistent with the expectation of image bi-level thresholding segmentation. To effectively granulate the image, a granule size selection method based on the homogeneity histogram is proposed, which is helpful in taking care of small objects and local variations of the images. The proposed thresholding method is evaluated by comparison with other three existing rough entropy based thresholding methods and three state-of-the-art image thresholding methods both qualitatively and quantitatively, using the natural images, the non-destructive testing images and an infrared video sequence from the OTCBVS Benchmark Data set. The comparison confirms the effectiveness of the proposed algorithm that not only is simple in form and clear in meaning, but also has good segmentation effect.

This paper aims to present a review of recent techniques in high dynamic range imaging (HDRI), which was the topic in research areas including image processing, computer graphics, and photography. HDRI or just HDR is a set of techniques that allows a greater dynamic range between the lightest and darkest areas of an image than current standard digital imaging techniques or photographic methods. HDR imaging technologies will spread its sphere of influence in imaging industry for its high quality and its powerful expression ability, including digital cinema, digital photography and next generation broadcast. This paper discusses recent advances, future direction in HDRI. The major goal of the paper is to provide a reference source for the researchers involved in HDRI, regardless of particular application areas.

Support vector machines (SVMs) have shown remarkable success in many applications. However, the non-smooth feature of objective function is a limitation in practical application of SVMs. To overcome this disadvantage, a twice continuously differentiable piecewise-smooth function is constructed to smooth the objective function of unconstrained support vector machine (SVM), and it issues a piecewise-smooth support vector machine (PWESSVM). Comparing to the other smooth approximation functions, the smooth precision has an obvious improvement. The theoretical analysis shows PWESSVM is globally convergent. Numerical results and comparisons demonstrate the classification performance of our algorithm is better than other competitive baselines.

Linear Discriminant Analysis (LDA) is a classical statistical approach for supervised feature extraction and dimensionality reduction, hard c-means (HCM) is a classical unsupervised learning algorithm for clustering. Based on the analysis of the relationship between LDA and HCM, Linear Discriminant Analysis-guided adaptive subspace hard c-means clustering algorithm (LDA-HCM) had been proposed. LDA-HCM combines LDA and HCM into a coherent framework and can adaptively reduce the dimension of data while performing data clustering simultaneously. Seeing that LDA-HCM is still a hard clustering algorithm, we consider the fuzzy extension version of LDA-HCM in this paper. To this end, firstly, we propose a new optimization criterion of Fuzzy Linear Discriminant Analysis (FLDA) by extending the value of membership function in classical LDA from binary 0 or 1 into closed interval [0, 1]. In the meantime, we present an efficient algorithm for the proposed FLDA. Secondly, we show the close relationship between FLDA and Maximum Entropy Fuzzy Clustering Algorithm (MEFCA): they both are maximizing fuzzy between-class scatter and minimizing within-class scatter simultaneously. Finally, based on the above analysis, combining FLDA and MEFCA into a joint framework, we propose fuzzy Linear Discriminant Analysis-guided maximum entropy fuzzy clustering algorithm (FLDA-MEFCA). LDA-MEFCA is a natural and effective fuzzy extension of LDA-HCM. Due to the introduction of soft decision strategy, FLDA-MEFCA can yield fuzzy partition of data set and is more flexible than LDA-HCM. We also give the convergence proof of FLDA-MEFCA. Extensive experiments on a collection of benchmark data sets are presented to show the effectiveness of the proposed algorithm.

Focusing on the problem of localized content-based image retrieval, based on probabilistic latent semantic analysis (PLSA) and transductive support vector machine (TSVM), a novel semi-supervised multi-instance learning (SSMIL) algorithm is proposed, where a bag corresponds to an image and an instance corresponds to the low-level visual features of a segmented region. In order to convert an MIL problem into a standard supervised learning problem, first, all the instances in training bags be clustered by K-Means method, and regards each cluster center as “visual-word” to build a visual vocabulary. Second, according to the distance between “visual-word” and instance, a fuzzy membership function is defined to establish a fuzzy term-document matrix, then use PLSA method to obtain bag’s (image’s) latent topic models, which can convert every bag to a single sample. Finally, in order to use the unlabeled images to improve retrieval accuracy, using semi-supervised TSVM to train classifiers. Experimental results on the COREL data sets show that the proposed method, named PLSA-SSMIL, is robust, and its performance is superior to other key existing MIL algorithms.

An improved Euclidean geometry approach to design quasi-cyclic (QC) Low-density parity-check (LDPC) codes with high-rate and low error floor is presented. The constructed QC-LDPC codes with high-rate have lower error floor than the original codes. The distribution of the minimum weight codeword is analyzed, and a sufficient existence condition of the minimum weight codeword is found. Simulations show that a lot of QC-LDPC codes with lower error floor can be designed by reducing the number of the minimum weight codewords satisfying this sufficient condition.

Otsu method is a commonly used technique for image thresholding selection, which can provide satisfactory results for thresholding an image with a histogram of clear bimodal distribution. This method, however, fails if the histogram is unimodal or close to unimodal. Ng (Ng, H.F., 2006. Automatic thresholding for defect detection. Pattern Recognition Lett. 27 1644-1649) proposed a modified method, called as valley-emphasis method, which weighs the objective function of the Otsu method with the valley point of the histogram for defect detection. In this paper, a revised valley-emphasis thresholding method is presented using the neighborhood information of the valley point. Experiment results show that the proposed method has the proper segmentation results in defect detection and provides better segmentation results than that of valley-emphasis method and Otsu method.

A novel network protocol, enhanced cooperative medium access control (ECoop MAC), is present in this article. Its function is to guarantee the quality of service (QoS) in wireless local area networks. For the sake of supporting different application scenarios, two proposed schemes, namely E-scheme I for lower priority traffic and E-scheme II for higher priority traffic can be adopted independently or in combination. ECoop MAC takes into account request failure problems, and utilizes cooperative protocol information to boost the system performance as well as to effectively cut control packets overhead. Simulation results show that the proposed algorithm can not only improve network throughput, but also lead to reduced network delays for individual packets.

Image encryption and hiding effect evaluation standard is used in image encryption and hiding algorithm widely. In order to evaluate the effect of image encryption and hiding fleetly and effectually, new image encryption and hiding effect evaluation indexes are presented based on probability of pixel's neighborhood. Simulation results show that the new method is simple and approach to man's subjective results. At the same time, a new image edge detection algorithm is presented using the evaluation indexes. Simulation results show that our method is simple and better effect can be got.

The design and analysis of multidimensional All-Partial-Sums (APS) algorithms are considered. We employ the sequence length as the performance measurement criterion for APS algorithms and corresponding thresholding methods, which is more sophisticated than asymptotic time complexity under the straight-line program computation model. With this criterion, we propose the piling algorithm to minimize the sequence length, then we show this algorithm is an optimal APS algorithm in commutative semigroups in the worst case. The experimental results also show the algorithmic efficiency of the piling algorithm. Furthermore, the theoretical works of APS algorithm will help to construct the higher dimensional thresholding methods.

Considering the important role of representation element in an interval-valued fuzzy set for decision analysis, the concepts of three parameters interval-valued fuzzy set and three parameters interval-valued fuzzy value are proposed, and some operators on three parameters interval-valued fuzzy values are given, and a new distance between three parameters interval-valued fuzzy values is defined. Based on the proposed distance, a TOPSIS decision-making method is shown for a multicriteria decision-making model on three parameters interval-valued fuzzy sets. Finally, a practical example is provided to illustrate the method's effectiveness.

Based on the defect of rival checked fuzzy c-means clustering algorithm, a new algorithm: suppressed fuzzy c-means clustering algorithm is proposed. The new algorithm overcomes the shortcomings of the original algorithm, establishes more natural and more reasonable relationships between hard c-means clustering algorithm and fuzzy c-means clustering algorithm.

The purpose of this paper is twofold. Firstly, a general conclusion about fuzzy entropy induced by distance measure is presented based on the axiom definitions of fuzzy entropy and distance measure. Secondly, some fuzzy entropy formulas which relate to the fuzzy entropy formula defined by De Luca and Termini (Inform. Control 20 (1972) 301) are given.

Based on the axiom definition of distance measure, some new formulas of fuzzy entropy induced by distance measure and some new properties of distance measure are given. A characterization of a σ-distance measure is a metric defined on all fuzzy sets is obtained.

New Hausdorff-like metrics for fuzzy sets are defined on a support set S in a metric (e.g., Euclidean) space. They modify the Hausdorff-like metrics given by Chaudhuri and Rosenfeld (1996) .

Intelligent video surveillance has found a wide range of applications in public security. This paper describes the state-of- the-art techniques in video surveillance system with single camera. This can serve as a starting point for building practical video surveillance systems in developing regions, leveraging existing ubiquitous infrastructure. In addition, this paper discusses the gap between existing technologies and the requirements in real-world scenario, and proposes potential solutions to reduce this gap.

Based on energy distribution analysis of orthogonal FRIT coefficients, a novel digital image watermark embedding and blind detecting algorithm in ridgelet domain is proposed in this paper. Since the ridgelet transform has directional sensitivity and anisotropy, the image is first partitioned into small pieces and the orthogonal FRIT is applied for each piece to obtain a sparse representation of the image, especially for straight edge singularity. Through analyzing texture distribution in ridgelet coefficients of each piece, these image pieces are classified into frat regions and texture regions by Fuzzy C-Mean ( FCM ) clustering algorithm . The texture regions used for watermark embedding also classified into two regions, strong texture regions and week texture regions for different embedding strength s based on the feature s of luminance masking and texture masking. And the watermarks can be blindly detected without the original image and watermark information . Experimental results show that the proposed watermarking scheme can achieve a better tradeoff between the robustness and the transparency.

For K-harmonic means(KHM) clustering algorithm and its generalized form: KHM P . clustering algorithm, fuzzy c-means clustering algorithm (FCM) and its generalized form: GFCM P clustering algorithms, the relations between KHM and FCM, KHM P and GFCM P are studied. By using the reformulation of the GFCM P , the facts that KHM P is a special case of FCM P as fuzzy parameter m is 2 and parameterp is greater than 2, and KHM is FCM as fuzzy parameter m is 2 are revealed. By using the theory of Robust Statistics, the performances of FCM P under different parameter p is studied and the conclusions are obtained: GFCM p is sensitive to noise when parameter p is greater than 1; it is robust to noise when p is less than 1. Experimental results show the correctness of our analysis.

Fuzzy number type data is a typical class of fuzzy data, and it can be regarded as a general form of the interval data and the crisp data. This paper studies fuzzy clustering algorithm for triangular fuzzy numbers. First of all, we give a novel distance between triangular fuzzy numbers by using three parameters interval number, and prove that the proposed distance is a complete metric on the set of triangular fuzzy numbers. And then, based on this novel distance, we propose two fuzzy c-means type clustering algorithms for dealing with triangular fuzzy numbers. Finally, some numerical examples are provided to illustrate the algorithm’s effectiveness.

Combining the human visual system with the image local properties, this paper proposes a novel watermarking algorithm based on ridgelet transform and Fuzzy C-Means (FCM) clustering. In order to obtain a sparse representation of the image, especially for straight edge singularity, the image is first partitioned into small pieces and the ridgelet transform is applied for each piece. The image pieces are adaptively classified into frat regions and texture regions by applying FCM clustering algorithm through analyzing texture distribution in ridgelet coefficients of each piece. And the watermarking is embedded into the middle ridgelet subband in the highest energy directions of texture pieces, where the embedding strength is also adjusted by the analysis of ridgelet coefficients of each piece based on the feature of luminance masking and texture masking. Then the embedded watermark can be blindly detected by correlation detector. Experimental results prove robustness and transparency of the proposed watermarking scheme.

This paper discusses a multi-criteria decision making model on intuitionistic fuzzy sets. Based on the similarity measure between intuitionistic fuzzy sets, a novel method is shown for the multi-criteria decision making model, the starting point of the proposed method is a geometrical interpretation of intuitionistic fuzzy set. An alternative is mapped to an intuitionistic fuzzy value by using the degree of similarity, and then a score function is used to measure the degree of suitability that an alternative satisfies the decision maker’s requirement. Examples are given to show the proposed method’s effectiveness.

Image thresholding method based on generalized fuzzy entropy segments the image using the principle that the membership degree of the threshold point is equal to m (0

A method is presented for three-level image segmentation through maximizing the fuzzy partition entropy of two-dimensional histogram. Two groups, each including three member functions, namely Z-function, i¾?-function and S-function, are used for fuzzy division of two-dimensional histogram to get nine fuzzy sets. And the nine fuzzy sets are classified to three parts, corresponding to dark, gray and white part of the image, respectively, while a fuzzy partition is obtained for the two-dimensional space. Then a fuzzy partition entropy is defined based on multi-dimensional fuzzy partition and entropy theory. The parameters of the six membership functions can be determined by maximizing fuzzy partition entropy of two-dimensional histogram and the procedure for finding the optimal combination of all the fuzzy parameters is implemented by quantum genetic algorithm with an appropriate coding method. The experiment results show that the proposed method gives better performance than onedimensional three-level thresholding method under noise case.

Traditional permutation encryption algorithm is not robustness for noise disturbing and shear transformation attacks. In order to ameliorate the security of image encryption algorithm, we present an image encryption algorithm based on location transformation. The algorithm permute image based on chaotic system and storage everyone pixel of the image in multi-place, this encrypted image is robustness for noise disturbing and shear transformation attack. An extended magic square matrix-generating algorithm is also presented and it improves on the efficiency of the magic square matrix-generating algorithm. The simulation results show that the effect of decrypting image is good when the encrypting image is modified by noise disturbing and shear transformation attack

Chaotic optimization is a new optimization technique. For image segmentation, conventional chaotic sequence is not very effective to three-dimension gray histogram. In order to solve this problem, a three-dimension chaotic sequence generating method is presented. Simulation results show that the generated sequence is pseudorandom and its distribution is approximately inside a sphere whose centre is (0.5 , 0.5 , 0.5). Based on this work, we use the proposed chaotic sequence to optimize three-dimension maximum between-variance image segmentation method. Experiments results show that our method has better segmentation effect and lower computation time than that of the original three-dimension maximum between-variance image segmentation method for mixed noise disturbed image.

Chaotic optimization is a new optimization technique. Conventional two-dimension chaotic sequence is not a good way to two-dimension gray histogram image segmentation because it is proportional distributing in [0,1] times [0,1]. In order to generate a better chaotic sequence that is fit to two- dimension gray histogram. A chaotic sequence generating method is proposed based on Arnold chaotic system and Bezier curve generating algorithm. The main feature of the new chaotic sequence is that its distribution is approximately inside a disc whose center is (0.5,0.5), this means that the sequence is superior to Arnold chaotic sequences in image segmenting. As application, a two-dimension maximum entropy image segmentation method is presented based on chaotic optimization. Simulation results show that our method has better segmentation effect and lower computation time than the original two-dimension maximum entropy method.

Linkens and Chen (1999) introduced a cluster validity function, which modified Bezdek's partition entropy. In this paper, according to their idea, a new cluster validity function, which modified Bezdek's partition coefficient, is defined. Some properties are studied and comparison experiments with partition coefficients are given.