Between-class Variance Research Articles

Design of Threshold Segmentation Method for Quantum Image

In this paper, a design method of quantum circuits for threshold segmentation of gray-scale image is proposed, which is a quantum counterpart of classical optimum global thresholding using Otsu’s method. First, we design a quantum circuit to obtain the histogram of the pixels distribution. Then, on this basis, we design the quantum circuits of some sub-modules, such as the cumulative histogram, cumulative mean, multiplication, division, the between-class variance and so on. Finally, all the between-class variances corresponding to each gray-level are sorted. At this time, the gray-level corresponding to the maximum of the between-class variance is the Otsu’s threshold. For all the operations involved in the classical Otsu’s threshold selection, we designed the corresponding quantum circuits, analyzed the computational complexity, and verified the correctness of these circuits by simulations on classical computers.

Relating Performance on Written Assessments to Features of Mathematics Discussion

Many researchers have illustrated the multi-faceted nature of classroom mathematics discussions, but few have demonstrated the effect of discussion on students’ assessment performance. We developed and employed a discussion observation instrument in 20 third and fourth grade classrooms in an economically disadvantaged, linguistically diverse school district and used hierarchical linear modeling (HLM) to determine whether between-class variation in word-problem-test scores can be explained by levels of class discussion. Results suggest overall class discussion scores, as well as two specific discussion features, variety of approaches and opportunities to speak, are significantly related to test performance. These results suggest classroom instruction including high-level math discussion may improve students’ performance on written measures of achievement.

A cuckoo search optimization-based forward consecutive mean excision model for threshold adaptation in cognitive radio

The forward consecutive mean excision (FCME) algorithm is one of the most effective adaptive threshold estimation algorithms presently deployed for threshold adaptation in cognitive radio (CR) systems. However, its effectiveness is often limited by the manual parameter tuning process and by the lack of prior knowledge pertaining to the actual noise distribution considered during the parameter modeling process of the algorithm. In this paper, we propose a new model that can automatically and accurately tune the parameters of the FCME algorithm based on a novel integration with the cuckoo search optimization (CSO) algorithm. Our model uses the between-class variance function of the Otsu’s algorithm as the objective function in the CSO algorithm in order to auto-tune the parameters of the FCME algorithm. We compared and selected the CSO algorithm based on its relatively better timing and accuracy performance compared to some other notable metaheuristics such as the particle swarm optimization, artificial bee colony (ABC), genetic algorithm, and the differential evolution (DE) algorithms. Following close performance values, our findings suggest that both the DE and ABC algorithms can be adopted as favorable substitutes for the CSO algorithm in our model. Further simulation results show that our model achieves reasonably lower probability of false alarm and higher probability of detection as compared to the baseline FCME algorithm under different noise-only and signal-plus-noise conditions. In addition, we compared our model with some other known autonomous methods with results demonstrating improved performance. Thus, based on our new model, users are relieved from the cumbersome process involved in manually tuning the parameters of the FCME algorithm; instead, this can be done accurately and automatically for the user by our model. Essentially, our model presents a fully blind signal detection system for use in CR and a generic platform deployable to convert other parameterized adaptive threshold algorithms into fully autonomous algorithms.

A multi-level image thresholding approach using Otsu based on the improved invasive weed optimization algorithm

Multi-level thresholding is one of the most popular techniques in image segmentation. However, selecting the optimal thresholds with high accuracy and efficiency is still challenging. In this paper, a novel multi-level thresholding method using between-class variance (Otsu) based on an improved invasive weed optimization algorithm (FIWO) is proposed. In the FIWO algorithm, the forking technique of the lightning search algorithm is introduced to guarantee the quality of the initial population and to enhance the exploration of the algorithm. In addition, the current best solution swing operation is used to obtain the optimal thresholds with a fast convergence rate. Comparative experiments are carried out to test the performance of FIWO. The results show that the proposed FIWO algorithm is able to achieve better segmented images with fewer iterations than those of the simulated annealing algorithm, gravitational search algorithm, whale optimization algorithm and traditional invasive weed optimization algorithm.

Combining Multifactorial Assessment Tools and Dimensionality Reduction Analysis for Fall Risk Classification in Community-Dwelling Older Adults

Background and Purpose: Optimal approaches in fall risk assessment involve interdisciplinary collaboration of assessment. This current work aimed at screening the fall risk characteristics from the objective balance and mobility tests between older fallers and nonfallers and further assessing the feasibility of 2 statistical dimensionality reduction models, Linear Discriminant Analysis (LDA) and Generalized Discriminant Analysis (GDA) for discriminating older nonspecific fallers. We hypothesized that the high-dimensionality objective sensor-based parameters, followed by a feature selection and dimensionality reduction process, would be able to discriminate older nonspecific fallers. Methods: Thirty-one community-living older individuals who were older than 60 years (faller: n = 15; nonfaller: n = 16) were recruited. The measurements include gait, balance, and ankle proprioception performances. LDA and GDA were further applied to obtain more discriminative feature space. Receiver-operating characteristic (ROC) curves were constructed to compare the classification quality in all the features. Results: Although some features in single objective measure reached statistical significance, the original features still resulted in high within-class and low between-class variances in the feature space. By further applying LDA and GDA on the original features, the performance of LDA in the feature space was improved. The area under the curve of ROC was GDA dimensionality reduction feature (1), LDA dimensionality reduction feature (0.99), proprioception (0.752), inertial measurement unit (0.745), and center of pressure (0.72), respectively. Conclusions: Experimental results showed the GDA feature has the best classification quality and the additional advantage in combination of interdisciplinary multifactorial fall risk assessment.

Image Segmentation Using Multilevel Thresholding: A Research Review

Image segmentation is a basic problem in computer vision and various image processing applications. Over the years, commonly used image segmentation has become quite challenging because of its utilization in many applications. Image thresholding is one of the most exploited techniques to accomplish image segmentation. Multilevel thresholding is found to be most appropriate and well known among all the image segmentation techniques. The segmented image quality is based on the techniques incorporated to choose the threshold value. In this paper, an exhaustive survey has been carried out, considering both the general purpose and satellite images to cover the performance comparison of various image segmentation approaches based on meta-heuristics optimization algorithms, present in the literature for multilevel image thresholding. In addition, this paper also focuses on information theoretic approach-based objective criterion using different statistical properties such as between-class variance, entropy, moment and maximum likelihood for selecting multilevel thresholds. A list of 157 publications on the subject is also appended for quick reference.

Analysis of true-color images from unmanned aerial vehicle to assess salinity stress on date palm

Soils in arid countries predominantly suffer from salinity and drought related to environmental problems that can lead to crop stress and low productivity. In this study, true-color aerial images from an unmanned aerial vehicle were used to assess the effect of soil and water salinity on date palm growth. Random soil samples ( n = 75) and irrigation water samples were collected from five sites and chemically analyzed. The custom algorithms were developed in ENVI® and MATLAB.® Green leaf index (GLI) was implemented to determine crop canopy attributes. Two segmentation methods namely between-class variance and foreground pixels were used to recognize the vegetation cover from other image pixels. The image analysis demonstrated that the mean value of GLI increased as the salinity levels decreased, R = 0.96 and 0.92 for soil electrical conductivity (EC) and water EC, respectively. The percentage of area covered with vegetation was correlated to soil EC and water EC with about 70% accuracy. On the other hand, the percentage of area covered with palm trees only was used accurately to evaluate the soil EC by R 2 = 0.89 and the water EC by R 2 = 0.86. The findings of this research can set foundations for the development of aerial color imaging on salinity stressed date palm monitoring, providing useful information for decision makers on salinity management.

Spatial context-based optimal multilevel energy curve thresholding for image segmentation using soft computing techniques

Image segmentation using multilevel thresholding (MT) is one of the leading methods. Although, as most techniques are based on the image histogram to be segmented, MT approaches only include the occurrence frequency of particular intensity range disregarding each spatial information. Energy curve-based contextual information can help to improve the quality of the thresholded image as it computes not only the value of the pixel but also its vicinity. Therefore, the energy curve is intended to carry spatial information into a curve with the same properties as the histogram. In this paper, classical Otsu’s method (between-class variance) is combined with energy curve for multilevel thresholding to perform segmentation of colored images. The energy curve-based Otsu (Energy-Otsu) uses an exhaustive search process to determine the optimal threshold values. However, due to the multidimensionality and multimodal nature of the color images, it becomes challenging and highly complex to obtain optimal thresholds. Therefore, the cuckoo search (CS) algorithm is coupled with Otsu thresholding criteria to perform MT over the energy curve. The proposed Energy-Otsu-CS method produces better-segmented results as compared to other well-known optimization algorithms such as differential evolution, particle swarm optimization, firefly algorithm, bacterial foraging optimization, and artificial bee colony algorithm. The proposed approach is examined intensively regarding quality, and the numerical parameter analysis is presented to compare the segmented results of the algorithms against closely related current approaches such as gray-level co-occurrence matrix and Renyi’ entropy-based thresholding approaches.

Learning personalized binary codes for finger vein recognition

This paper aims to develop a new local binary learning feature for finger vein images called personalized binary code (PBC). Existing local learning features of the image, such as compact binary face descriptor (CBFD) and discriminant binary descriptor (DBD), ignore the structure of the binary feature corresponding to each class when learning. Inspired by the bit-consistency phenomenon of local binary pattern (LBP), we combine the constraints of within-class sparse constraints with linear discriminant analysis (LDA) criteria in the learning process to explore the bit-consistent structure of learned binary features. Specifically, a multi-directional pixel difference vector is first calculated at each pixel of the training set of finger vein images, and an objective function is designed to learn a discriminative map that projects these multi-directional pixel difference vectors into binary features. Our objective function consists of several constraints: Minimizing quantization error, minimizing within-class variance of learned binary features, and maximizing the between-class variance of learned binary features, and minimizing the L2,1 norm of the learned features for each class. Next, the algorithm based on alternating direction method (ADM) is designed to solve the optimization problem. Finally, a histogram representation of each finger vein image is constructed by the bag-of-words model. In our experiments, the proposed PBC shows that the leading performance on two public databases in comparison with existing finger vein recognition methods.

Multi-level image thresholding based on social spider algorithm for global optimization

Thresholding is one of the simplest and popular technique for segmenting images. Maximum between-class variance (Otsu’s) method is one of the well-known and widely used method in case of segmentation. Not only Otsu could be used for bi-level thresholding but also it could be extended to multi-level image thresholding. Finding the optimum threshold values in multi-level case is very time consuming process, thus optimization algorithm can deal with this problem. In this paper social spider algorithm for global optimization has been used for maximizing the between-class variance to carry out multi-level image thresholding. Experimental outcomes have demonstrated that the proposed method is capable of estimating threshold values and yield satisfying outcome.

An improved background segmentation algorithm for fringe projection profilometry based on Otsu method

In fringe projection profilometry, there always exist background and shadow areas in the captured fringe images. To improve the data quality, these areas should be recognized and deleted before the point cloud is reconstructed. For the existing algorithms, it is a nontrivial task to obtain a reasonable threshold to recognize the background and shadow areas from the fringe images. An inappropriate threshold usually leads to a misclassification of the object and background/shadow areas. In this paper, an improved adaptive thresholding method is built to address this problem. Instead of the gray level histogram, the modulation level histogram is utilized to compute the threshold. The modulation image is computed from all the captured images in the projection sequence. A novel objective function with a Gaussian weighted factor is proposed to improve Otsu’s method. With this weighted factor, the between-class variance between the object and background/shadow is intensified. The new built objective function also makes the objective function plot much smoother, which makes the proposed method more suitable for noise image segmentation. Additionally, the weighted factor changes adaptively with the input images. Experiment results illustrate that the proposed scheme performs better than existing methods. The scan for an area on a car body rear demonstrates the effectiveness and feasibility of the proposed method in structured light measurement of real metal components.

Mask-based denoising scheme for ghost imaging**Project supported by the National Natural Science Foundation of China (Grant No. 61627823), the Foundation for Excellent Young Talents of Jilin Province, China (Grant No. 20190103010JH), the “13th Five-Year” Science and Technology Research Project of the Education Department of Jilin Province, China (Grant No. JJKH20190277KJ), the China Postdoctoral Science Foundation (Grant No. 2018M641759), and the

Ghost imaging (GI) is thought of as a promising imaging method in many areas. However, the main drawback of GI is the huge measurement data and low signal-to-noise ratio. In this paper, we propose a novel mask-based denoising scheme to improve the reconstruction quality of GI. We first design a mask through the maximum between-class variance (OTSU) method and construct the measurement matrix with speckle patterns. Then, the correlated noise in GI can be effectively suppressed by employing the mask. From the simulation and experimental results, we can conclude that our method has the ability to improve the imaging quality compared with traditional GI method.

Multisource Region Attention Network for Fine-Grained Object Recognition in Remote Sensing Imagery

Fine-grained object recognition concerns the identification of the type of an object among a large number of closely related sub-categories. Multisource data analysis, that aims to leverage the complementary spectral, spatial, and structural information embedded in different sources, is a promising direction towards solving the fine-grained recognition problem that involves low between-class variance, small training set sizes for rare classes, and class imbalance. However, the common assumption of co-registered sources may not hold at the pixel level for small objects of interest. We present a novel methodology that aims to simultaneously learn the alignment of multisource data and the classification model in a unified framework. The proposed method involves a multisource region attention network that computes per-source feature representations, assigns attention scores to candidate regions sampled around the expected object locations by using these representations, and classifies the objects by using an attention-driven multisource representation that combines the feature representations and the attention scores from all sources. All components of the model are realized using deep neural networks and are learned in an end-to-end fashion. Experiments using RGB, multispectral, and LiDAR elevation data for classification of street trees showed that our approach achieved 64.2% and 47.3% accuracies for the 18-class and 40-class settings, respectively, which correspond to 13% and 14.3% improvement relative to the commonly used feature concatenation approach from multiple sources.

A context sensitive energy thresholding based 3D Otsu function for image segmentation using human learning optimization

In this paper, a novel context-based 3D Otsu algorithm using human learning optimization (HLO) is proposed for multilevel color image segmentation. The performance of 3D Otsu algorithm is reported to be poor while dealing with between-class variances through the aid of three-dimensional histogram. To overcome this problem, the concept of context thresholding has been exploited to derive pixel intensity values and spatial information. The nature of spatial context and histogram of an image is very similar. The use of energy curve for 3D Otsu gives satisfactory results but it is more time-consuming during the process of threshold selection. HLO is a recently developed meta-heuristic optimization algorithm that involves the use of learning operators developed by mimicking human learning mechanisms. In this paper, in order to avoid an exhaustive search to obtain optimal thresholds, HLO is used. Experimental studies reported in this paper demonstrate that the proposed method is better than the histogram-based 1D Otsu, 2D Otsu, and 3D Otsu methods. These claims have been confirmed by comparing fidelity parameters such as mean error (ME), mean squared error (MSE), peak signal-to-noise ratio (PSNR), feature similarity index (FSIM), structure similarity index (SSIM) and entropy.

Dynamic Harris Hawks Optimization with Mutation Mechanism for Satellite Image Segmentation

In this paper, a novel satellite image segmentation technique based on dynamic Harris hawks optimization with a mutation mechanism (DHHO/M) is proposed. Compared with the original Harris hawks optimization (HHO), the dynamic control parameter strategy and mutation operator used in DHHO/M can avoid falling into the local optimum and efficiently enhance the search capability. To evaluate the performance of the proposed method, a series of experiments are carried out on various satellite images. Eight advanced thresholding approaches are selected for comparison. Three criteria are adopted to determine the segmentation thresholds, namely Kapur’s entropy, Tsallis entropy, and Otsu between-class variance. Furthermore, four oil pollution images are used to further assess the practicality and feasibility of the proposed method on real engineering problem. The experimental results illustrate that the DHHO/M based thresholding technique is superior to others in the following three aspects: fitness function evaluation, image segmentation effect, and statistical tests.

Towards fast and kernelized orthogonal discriminant analysis on person re-identification

Recognizing a person across different non-overlapping camera views, is the task of person re-identification. For achieving the task, an effective way is to learn a discriminative metric by minimizing the within-class variance and maximizing the between-class variance simultaneously. However, the dimension of sample feature vector is usually greater than the number of training samples, as a result, the within-class scatter matrix is singular and the metric cannot be learned. In this paper, we propose to solve the singularity problem by employing the pseudo-inverse of the within-class scatter matrix and learning an orthogonal transformation for the metric. The proposed method can be effectively solved with a closed-form solution and no parameters required to tune. In addition, we develop a kernel version against non-linearity in person re-identification, and a fast version for more efficient solution. In experiments, we prove the validity and advantage of the proposed method for solving the singularity problem in person re-identification, and analyze the effectiveness of both kernel version and fast version. Extensively comparative experiments on VIPeR, PRID2011, CUHK01 and CUHK03 person re-identification benchmark datasets, show the state-of-the-art results of the proposed method.

Kapur’s entropy based optimal multilevel image segmentation using Crow Search Algorithm

Image segmentation is an essential part of image analysis, which has a direct impact on the quality of image analysis results. Thresholding is one of the simplest and widely used methods for image segmentation. Thresholding can be either bi-level, which involves partitioning of an image into two segments, or multilevel, which partitions an image into multiple segments using multiple thresholds values. This paper focuses on multilevel thresholding. A good segmentation scheme through multilevel thresholding identifies suitable threshold values to optimize between-class variance or entropy criterion. For such optimizations, nature inspired metaheuristic algorithms are commonly used. This paper presents a Kapur’s entropy based Crow Search Algorithm (CSA) to estimate optimal values of multilevel thresholds. Crow Search Algorithm is based on the intelligent behavior of crow flock. Crow Search Algorithm have shown better results because of less number of parameters, no premature convergence, and better exploration–exploitation balance in the search strategy. Kapur’s entropy is used as an objective function during the optimization process. The experiments have been performed on benchmarked images for different threshold values (i.e. 2, 4, 8, 16, 32 thresholds). The proposed method has been assessed and performance is compared with well-known metaheuristic optimization methods like Particle Swarm Optimization (PSO), Differential Evolution (DE), Grey Wolf Optimizer (GWO), Moth-Flame Optimization (MFO) and Cuckoo Search (CS). Experimental results have been evaluated qualitatively and quantitatively by using well-performed evaluation methods namely PSNR, SSIM, and FSIM. Computational time and Wilcoxon p-type value also compared. Experimental results show that proposed algorithm performed better than PSO, DE, GWO, MFO and CS in terms of quality and consistency.

Dragonfly Algorithm with Opposition-Based Learning for Multilevel Thresholding Color Image Segmentation

Multilevel thresholding is a very active research field in image segmentation, and has been successfully used in various applications. However, the computational time will increase exponentially as the number of thresholds increases, and for color images which contain more information this is even worse. To overcome the drawback while maintaining segmentation accuracy, a modified version of dragonfly algorithm (DA) with opposition-based learning (OBLDA) for color image segmentation is proposed in this paper. The opposition-based learning (OBL) strategy simultaneously considers the current solution and the opposite solution, which are symmetrical in the search space. With the introduction of OBL, the proposed algorithm has a faster convergence speed and more balanced exploration–exploitation compared with the original DA. In order to clearly demonstrate the outstanding performance of the OBLDA, the proposed method is compared with seven state-of-the-art meta-heuristic algorithms, through experiments on 10 test images. The optimal threshold values are calculated by the maximization of between-class variance and Kapur’s entropy. Meanwhile, some indicators, including peak signal to noise ratio (PSNR), feature similarity index (FSIM), structure similarity index (SSIM), the average fitness values, standard deviation (STD), and computation time are used as evaluation criteria in the experiments. The promising results reveal that proposed method has the advantages of high accuracy and remarkable stability. Wilcoxon’s rank sum test and Friedman test are also performed to verify the superiority of OBLDA in a statistical way. Furthermore, various satellite images are also included for robustness testing. In conclusion, the OBLDA algorithm is a feasible and effective method for multilevel thresholding color image segmentation.

Oppositional elephant herding optimization with dynamic Cauchy mutation for multilevel image thresholding

This paper presents an improved elephant herding optimization (IEHO) to solve the multilevel image thresholding problem for image segmentation by introducing oppositional-based learning (OBL) and dynamic cauchy mutation (DCM). OBL accelerates the convergence rate and enhances the performance of standard EHO whereas DCM mitigates the premature convergence. The suggested optimization approach maximizes two popular objective functions: ‘Kapur’s entropy’ and ‘between-class variance’ to estimate optimized threshold values for segmentation of the image. The performance of the proposed technique is verified on a set of test images taken from the benchmark Berkeley segmentation dataset. The results are analyzed and compared with conventional EHO and other four popular recent metaheuristic algorithms namely cuckoo search, artificial bee colony, bat algorithm, particle swarm optimization and one classical method named dynamic programming found from the literature. Experimental results show that the proposed IEHO provides promising performance compared to other methods in view of optimized fitness value, peak signal-to-noise ratio, structure similarity index and feature similarity index. The suggested algorithm also has better convergence than the other methods taken into consideration.

Image segmentation using multilevel thresholding based on modified bird mating optimization

Multilevel thresholding using Otsu or Kapur methods is widely used in the context of image segmentation. These methods select optimal thresholds in gray level images by maximizing between-class variance or entropy criterion. These methods become time consuming and less efficient with increasing number of thresholds. To increase the efficiency of the image segmentation using multilevel thresholding based on Kapur and Otsu methods, we developed a hybrid optimization algorithm named BMO-DE based on bird mating optimization (BMO) and differential evolutionary (DE) algorithms. The efficiency of the proposed method was evaluated on eight standard benchmark images. The proposed method achieved better segmentation results in term of solution quality and stability in comparison with other well-known techniques including bacterial foraging (BF), modified bacterial foraging (MBF), particle swarm optimization (PSO), genetic algorithm (GA) and hybrid algorithm named PSO-DE.