Change detection is a concerned task in the interpretation of remote sensing images. The mining of the relationship on change features is usually implicit in the deep learning networks that contain single-branch or two-branch encoders. However, due to the lack of artificial prior design for the relationship on change features, these networks cannot learn enough semantic information on change features and lead to the poor performance. So, we propose a new network architecture explicit change-relation network (ECRNet) for the explicit mining of change-relation features. In our study of the literature, our suggestion is that the change features for change detection should be divided into prechanged image features, postchanged image features, and change-relation features. In order to fully mining these three kinds of change features, we propose the triple branch network combining the transformer and convolutional neural network (CNN) to extract and fuse these change features from two perspectives of global information and local information, respectively. In addition, we design the continuous change-relation (CCR) branch to further obtain the continuous and detailed change-relation features to improve the change discrimination capability of the model. The experimental results show that our network performs better than those of the existing advanced networks by the F1 score improvements of 0.66/0.37/0.70/1.09 on the very high-resolution (VHR) remote sensing datasets of the LEVIR-CD/SVCD/WHU-CD/SYSU-CD. Our source code is available athttps://github.com/DalongZ/ECRNet.

Accurate volumetric segmentation of brain tumors and tissues is beneficial for quantitative brain analysis and brain disease identification in multi-modal Magnetic Resonance (MR) images. Nevertheless, due to the complex relationship between modalities, 3D Fully Convolutional Networks (3D FCNs) using simple multi-modal fusion strategies hardly learn the complex and nonlinear complementary information between modalities. Meanwhile, the indiscriminative feature aggregation between low-level and high-level features easily causes volumetric feature misalignment in 3D FCNs. On the other hand, the 3D convolution operations of 3D FCNs are excellent at modeling local relations but typically inefficient at capturing global relations between distant regions in volumetric images. To tackle these issues, we propose an Aligned Cross-Modality Interaction Network (ACMINet) for segmenting the regions of brain tumors and tissues from MR images. In this network, the cross-modality feature interaction module is first designed to adaptively and efficiently fuse and refine multi-modal features. Secondly, the volumetric feature alignment module is developed for dynamically aligning low-level and high-level features by the learnable volumetric feature deformation field. Thirdly, we propose the volumetric dual interaction graph reasoning module for graph-based global context modeling in spatial and channel dimensions. Our proposed method is applied to brain glioma, vestibular schwannoma, and brain tissue segmentation tasks, and we performed extensive experiments on BraTS2018, BraTS2020, Vestibular Schwannoma, and iSeg-2017 datasets. Experimental results show that ACMINet achieves state-of-the-art segmentation performance on all four benchmark datasets and obtains the highest DSC score of hard-segmented enhanced tumor region on the validation leaderboard of the BraTS2020 challenge.

The problem of path planning is a challenging task for mobile robots. A practical example can be seen in the robots commonly employed in warehouses: they must navigate to pick up goods and move them to certain locations. Therefore, the robot needs a method of moving from an initial location in the warehouse to a final location repeatedly. In this paper, we propose a technique that allows a robot to path plan in generalized environments, from different starting and goal locations. The method is based on a graph representation of the environment, and is capable of finding the shortest path between two points in the environment. In order to generalize appropriately, we show that a neural network is able to effectively choose the correct actions to take at each time step in the path planning problem.

Recently, with the advent of the Internet of everything and 5G network, the amount of data generated by various edge scenarios such as autonomous vehicles, smart industry, 4K/8K, virtual reality (VR), augmented reality (AR), etc., has greatly exploded. All these trends significantly brought real-time, hardware dependence, low power consumption, and security requirements to the facilities, and rapidly popularized edge computing. Meanwhile, artificial intelligence (AI) workloads also changed the computing paradigm from cloud services to mobile applications dramatically. Different from wide deployment and sufficient study of AI in the cloud or mobile platforms, AI workload performance and their resource impact on edges have not been well understood yet. There lacks an in-depth analysis and comparison of their advantages, limitations, performance, and resource consumptions in an edge environment. In this paper, we perform a comprehensive study of representative AI workloads on edge platforms. We first conduct a summary of modern edge hardware and popular AI workloads. Then we quantitatively evaluate three categories (i.e., classification, image-to-image, and segmentation) of the most popular and widely used AI applications in realistic edge environments based on Raspberry Pi, Nvidia TX2, etc. We find that interaction between hardware and neural network models incurs non-negligible impact and overhead on AI workloads at edges. Our experiments show that performance variation and difference in resource footprint limit availability of certain types of workloads and their algorithms for edge platforms, and users need to select appropriate workload, model, and algorithm based on requirements and characteristics of edge environments.

Brain tissue segmentation in multi-modal magnetic resonance (MR) images is significant for the clinical diagnosis of brain diseases. Due to blurred boundaries, low contrast, and intricate anatomical relationships between brain tissue regions, automatic brain tissue segmentation without prior knowledge is still challenging. This paper presents a novel 3D fully convolutional network (FCN) for brain tissue segmentation, called APRNet. In this network, we first propose a 3D anisotropic pyramidal convolutional reversible residual sequence (3DAPC-RRS) module to integrate the intra-slice information with the inter-slice information without significant memory consumption; secondly, we design a multi-modal cross-dimension attention (MCDA) module to automatically capture the effective information in each dimension of multi-modal images; then, we apply 3DAPC-RRS modules and MCDA modules to a 3D FCN with multiple encoded streams and one decoded stream for constituting the overall architecture of APRNet. We evaluated APRNet on two benchmark challenges, namely MRBrainS13 and iSeg-2017. The experimental results show that APRNet yields state-of-the-art segmentation results on both benchmark challenge datasets and achieves the best segmentation performance on the cerebrospinal fluid region. Compared with other methods, our proposed approach exploits the complementary information of different modalities to segment brain tissue regions in both adult and infant MR images, and it achieves the average Dice coefficient of 87.22% and 93.03% on the MRBrainS13 and iSeg-2017 testing data, respectively. The proposed method is beneficial for quantitative brain analysis in the clinical study, and our code is made publicly available.

Advanced robotic devices have the potential to improve both clinical and home-based rehabilitation procedures in stroke therapy. Having an active, intelligent device that can interact with the patient in both actuation and sensing feedback from the body would help improve the assessment of rehabilitation. Reliable signal detection and recognition of user intents are the key points of developing active robotic devices. Surface Electromyography (sEMG) technique is commonly used for non-invasive biological signal detection from muscle activations. This work presents a simple Convolutional Neural Network (CNN) model combined with A2C actor-critic algorithm-based reinforcement learning to predict simultaneous wrist motion intention direction. The proposed model was tested with experimental 2-channel sEMG datasets using both deep features extracted from CNN and hand-crafted features. We achieved an average accuracy of approximately 92% regardless of the instantaneous angular position of the wrist. We also presented generalization test results to demonstrate the performance of the model to a completely new subject's sEMG data.

Feature selection and construction are important pre-processing techniques in data mining. They allow not only dimensionality reduction but also classification accuracy and efficiency improvement. While feature selection consists in selecting a subset of relevant features from the original feature set, feature construction corresponds to the generation of new high-level features, called constructed features, where each one of them is a combination of a subset of original features. However, different features can have different abilities to distinguish different classes. Therefore, it may be more difficult to construct a better discriminating feature when combining features that are relevant to different classes. Based on these definitions, feature construction could be seen as a BLOP (Bi-Level optimization Problem) where the feature subset should be defined in the upper level and the feature construction is applied in the lower level by performing mutliple followers, each of which generates a set class dependent constructed features. In this paper, we propose a new bi-level evolutionary approach for feature construction called BCDFC that constructs multiple features which focuses on distinguishing one class from other classes using Genetic Programming (GP). A detailed experimental study has been conducted on six high-dimensional datasets. The statistical analysis of the obtained results shows the competitiveness and the outperformance of our bi-level feature construction approach with respect to many state-of-art algorithms.

Classification of high-dimensional hyperspectral data is important in the remote sensing community. Although there are many supervised and unsupervised models for classification, a subset of them are consistent in reproducibility with high confidence intervals. This study shows the consistency and model accuracy of Self-Organizing map (SOM) and Cellular Self-Organizing Map (CSOM) over the benchmark pattern and hyperspectral datasets. A constructive analysis of both the models and optimal use cases are discussed in this work. t-SNE visualizations of these models were added to understand the models capability of capturing the topological density of the original data. In this study, CSOM is adopted for remote sensing image classification as an unsupervised clustering algorithm. The primary difference between SOM and CSOM is the updating principle of the neighboring neurons. Our experimental results show that CSOM performs better than SOM for the overlapping classes while SOM is efficient for classes which are well-separated or separable.

Classification of remotely sensed images is a challenging task due to their inherently high dimensionality. Conventional methods to combat this issue involves using feature selection or extraction before feeding data to a discriminator. An intuitive approach known as Automatic Variable Weighting K-Means (W-K-means) incorporates the use of learned feature weights to the K - Means clustering algorithm to place emphasis on more prominent features. The inclusion of feature weights assists in the discovery of optimal cluster centers, thus increasing classification accuracy. As W-K-means was proposed for high-dimensional data, it is possible to achieve excellent hyperspectral image segmentation. However, the effectiveness in a high-dimensional setting was not thoroughly explored as the original experiments used datasets of low to medium dimensions. By combining feature extraction with W-K-means, essential features can be used to influence clustering. The experimental results show that Principal Component Analysis (PCA) with W-K-means performs exceptionally well in high-dimensional space when compared to W-K-means solely.

The multi‐atlas‐based methods are widely applied in the automatic labelling in magnetic resonance (MR) images. However, most multi‐atlas‐based methods require that all atlases be registered to the target image accurately to have a correct label propagation. In this study, the authors introduce the term spatial indexes and construct a hybrid atlas forest model to gather the labelling information from all atlases without propagating labels from every single atlas. Furthermore, a new automatic labelling method using the hybrid atlas forest model based on spatial indexes is proposed. In the proposed framework, an atlas is chosen arbitrarily as a reference image and the spatial indexes are constructed on this image space. Then, the samples are selected from all atlases in the dataset based on the spatial indexes to construct a samples pool. Finally, the hybrid atlas forest model will be trained on the samples pool and used to predict the labelling of the target. Experiments are conducted on two public datasets to evaluate the effectiveness of the proposed method. The experimental results show that the proposed method reduces the requirement of strong dependence on precise registration and improve the accuracy of labelling.

Early detection of lung cancer is an effective way to improve the survival rate of patients. It is a critical step to have accurate detection of lung nodules in computed tomography (CT) images for the diagnosis of lung cancer. However, due to the heterogeneity of the lung nodules and the complexity of the surrounding environment, it is a challenge to develop a robust nodule detection method. In this study, we propose a two-stage convolutional neural networks (TSCNN) for lung nodule detection. The first stage based on the improved U-Net segmentation network is to establish an initial detection of lung nodules. During this stage, in order to obtain a high recall rate without introducing excessive false positive nodules, we propose a new sampling strategy for training. Simultaneously, a two-phase prediction method is also proposed in this stage. The second stage in the TSCNN architecture based on the proposed dual pooling structure is built into three 3D-CNN classification networks for false positive reduction. Since the network training requires a significant amount of training data, we designed a random mask as the data augmentation method in this study. Furthermore, we have improved the generalization ability of the false positive reduction model by means of ensemble learning. We verified the proposed architecture on the LUNA dataset in our experiments, which showed that the proposed TSCNN architecture did obtain competitive detection performance.

Bi-clustering is one of the main tasks in data mining with several application domains. It consists in partitioning a data set based on both rows and columns simultaneously. One of the main difficulties in bi-clustering is the issue of finding the number of bi-clusters, which is usually a user-specified parameter. Recently, in 2017, a new multi-objective evolutionary clustering algorithm, called MOCK-II, has shown its effectiveness in data clustering while automatically determining the number of clusters. Motivated by the promising results of MOCK-II, we propose in this paper a hybrid extension of this algorithm for the case of bi-clustering. Our new algorithm, called MOBICK, uses an efficient solution encoding, an effective crossover operator, and a heuristic mutation strategy. Similarly to MOCK-II, MOBICK is able to find automatically the number of bi-clusters. The outperformance of our algorithm is shown on a set of real gene expression data sets against several existing state-of-the-art works. Moreover, to be able to compare MOBICK to MOCK-I and MOCK-II, we have designed two basic extensions of MOCK-I and MOCK-II for the case of bi-clustering that we named B-MOCK-I and B-MOCK-II. Again, the experimental results confirm the merits of our proposal.

Feature selection and construction are important pre-processing techniques in machine learning and data mining. They may allow not only dimensionality reduction but also classifier accuracy and efficiency improvement. Feature selection aims at selecting relevant features from the original feature set, which could be less informative to achieve good performance. Feature construction may work well as it creates new highlevel features, but these features do not have the same degree of importance, which makes the use of weighted-features construction a very challenging topic. In this paper, we propose a bi-level evolutionary approach for efficient feature selection and simultaneous feature construction and feature weighting, called Bi-level Weighted-Features Construction (BWFC). The basic idea of our BWFC is to exploit the bi-level model for performing feature selection and weighted-features construction with the aim of finding an optimal subset of features combinations. Our approach has been assessed on six high-dimensional datasets and compared against three existing approaches, using three different classifiers for accuracy evaluation. Experimental results show that our proposed algorithm gives competitive and better results with respect to the state-of-the-art algorithms.

Classification of imbalanced data is a challenging issue in the interpretation of remote sensing images. In a majority class (negatives), there are much more pixels than in a minority class (positives). This imbalanced data makes the classification extremely difficult to produce higher accuracy. The sampling technique is one of techniques, which work well for many pattern data, but not for remote sensing images. The Fast Box algorithm (FBA) which uses the clustering algorithm was then proposed to characterize and discriminate the minority class from the majority class by determining the decision boundaries of the minority data. The FBA performs well in most pattern datasets; however, it fails to recognize the minority class properly in remote sensing images. In this study, we propose a new Constrained Box Algorithm (CBA), which can effectively detect the minority class within remote sensing images. The FBA often misclassifies negatives as positives, CBA eliminates this issue by restricting the maximum number of allowed positives within a box. Our new algorithm finds the minority class by an iterative process of discovering appropriate boundaries using clustering and eliminating majority instances from initial boundaries. A threshold is used to guide the search process to find acceptable boundaries. The set of accepted boundaries are then used to discover the minority class. Experimental results demonstrates that the minority class was correctly detected in satellite images.

Feature selection and construction are important pre-processing techniques in data mining. They may allow not only dimensionality reduction but also classifier accuracy and efficiency improvement. These two techniques are of great importance especially for the case of high-dimensional data. Feature construction for high-dimensional data is still a very challenging topic. This can be explained by the large search space of feature combinations, whose size is a function of the number of features. Recently, researchers have used Genetic Programming (GP) for feature construction and the obtained results were promising. Unfortunately, the wrapper evaluation of each feature subset, where a feature can be constructed by a combination of features, is computationally intensive since such evaluation requires running the classifier on the data sets. Motivated by this observation, we propose, in this paper, a hybrid multiobjective evolutionary approach for efficient feature construction and selection. Our approach uses two filter objectives and one wrapper objective corresponding to the accuracy. In fact, the whole population is evaluated using two filter objectives. However, only non-dominated (best) feature subsets are improved using an indicator-based local search that optimizes the three objectives simultaneously. Our approach has been assessed on six high-dimensional datasets and compared with two existing prominent GP approaches, using three different classifiers for accuracy evaluation. Based on the obtained results, our approach is shown to provide competitive and better results compared with two competitor GP algorithms tested in this study.

To obtain a higher accuracy in the multi‐atlas patch‐based label fusion method, it is essential to have the accurate similarity measure of selected patches. In this study, the authors propose a new sparse patch‐based representation method using a local binary texture (LBT) in the atlas image and atlas label information for the multi‐atlas label fusion. In the proposed method, the intensity information in a patch is converted into a LBT which is then combined with the labels of corresponding patches from the atlas to form an atom of a dictionary. The initial labels of target images are estimated through a rough segmentation. The voxel in a patch to be labelled is also constructed as a vector similar to the atom. The voxel vector is then modelled as a sparse linear combination of the atoms in the dictionary. Experimental results on two MR brain data sets demonstrated that the proposed method is efficient in the segmentation which can achieve competitive performance compared with the state‐of‐the‐art methods.

Impulse noise detection is important to the restoration of color images contaminated by impulse noise in switching vector median filters. To increase detection accuracy, an effective color-impulse detector is presented. A new color distance metric based on quaternion theory is proposed. The proposed color distance metric is used to calculate the local density of a color pixel. A hard thresholding strategy is used to determine whether a color pixel is corrupted by impulse noise or not (i.e., an outlier). The noisy pixels detected will be restored by a weighted vector median filter, while the noise-free pixels remain unchanged. The experimental comparisons show that the proposed algorithm can obtain lower false and miss detection rate, and produces better performance in terms of peak signal-to-noise ratio and feature similarity measures, compared to other well-known color image filtering methods.

Spectral Unmixing is a challenging and absorbing problem. Unmixning allows us to break down a pixel's composition into its material components. Many avenues of spectral unmixing have been attempted with considerable success. One such avenue is to frame the spectral unmixing problem as an Estimation-Measurement problem and avail the use of the well-known Kalman Filter (KF) technique. Two such recent work has been the KF based Linear Unmixing (KFLU) approach and the KF approach for Hyperspectral Signature Estimation, Identification and Abundance Quantification (KFHSE/I/AQ). The above techniques aim to address the spectral unmixing and the spectral signature identification problems respectively. This work extends the above formulation by the use of the Particle Filter (PF) based filtering approach. The particle filter is a recent development in the KF framework. It addresses two major improvements over the KF. It enables use of nonlinearity in the estimation process and further allows fusion of multiple information sources. Additionally, by the use of distributed setup using particles, measurement errors are more efficiently reduced. The above enhancements are the primary motivation to create the proposed Kalman Particle Filter (KPF) in this paper. A major disadvantage in the use of Kalman Filter is the selection of the estimation matrix which can be efficiently resolved using the Kalman Particle Filter as described in this paper. Experiments performed using this new algorithm demonstrate the utility of the proposed approach as a better new tool to solve the spectral unmixing problem as compared to prior Kalman linear unmixing approach.

The multi‐Atlas patch‐based label fusion method (MAS‐PBM) has emerged as a promising technique for the magnetic resonance imaging (MRI) image segmentation. The state‐of‐the‐art MAS‐PBM approach measures the patch similarity between the target image and each atlas image using the features extracted from images intensity only. It is well known that each atlas consists of both MRI image and labelled image (which is also called the map). In other words, the map information is not used in calculating the similarity in the existing MAS‐PBM. To improve the segmentation result, the authors propose an enhanced MAS‐PBM in which the maps will be used for similarity measure. The first component of the proposed method is that an initial segmentation result (i.e. an appropriate map for the target) is obtained by using either the non‐local‐patch‐based label fusion method (NPBM) or the sparse patch‐based label fusion method (SPBM) based on the grey scales of patches. Then, the SPBM is applied again to obtain the finer segmentation based on the labels of patches. The authors called these two versions of the proposed fusion method as MAS‐PBM‐NPBM and MAS‐PBM‐SPBM. Experimental results show that more accurate segmentation results are achieved compared with those of the majority voting, NPBM, SPBM, STEPS and the hierarchical multi‐atlas label fusion with multi‐scale feature representation and label‐specific patch partition.

Noise detection and suppression is one of the important issues in digital image processing. In this study, we develop a new algorithm for detection and suppression of the impulse noise in remote sensing images. The algorithm, called Moran's I Spatial Autocorrelation filter (MSAF), is based on the Standard Median Filter and Moran's I which is used to measure the spatial autocorrelation. Our experimental results show that the MSAF has improved outcomes in terms of Peak Signal-to- Noise Ratio (PSNR) and Mean Square Error (MSE) compared to other filtering algorithms including the Standard Median Filter (SMF), Center Weighted Median Filter (CWMF), Adaptive Center Weighted Median Filter (ACWMF), Decision Based Filter (DBF), Signal-Dependent Rank Ordered Mean Filter (SDROMF) and Modified Decision Based Unsymmetric Trimmed Median Filter (MDBUTMF).

Fuzzy clustering is a widely used approach for data classification by using the fuzzy set theory. The probability measure and the possibility measure are two popular measures which have been used in the fuzzy [Formula: see text]-means algorithm (FCM) and the possibilistic clustering algorithms (PCAs), respectively. However, the numerical experiments revealed that FCM and its derivatives lack the intuitive concept of degree of belongingness, and PCAs suffer from the “coincident problem” and cannot provide very stable results for some data sets. In this study, we propose a new clustering algorithm, called the credibilistic clustering algorithm (CCA), based on the credibility measure. The credibility measure provides some unique properties which can solve the “coincident problem” and noise issue compared with the probability measure and possibility measure. Based on some randomly generated data sets, experimental results compared with FCM and PCA show that CCA can deal with the “coincident problem” with good clustering results, and it is more robust to noise than PCA.

A Quantum-Modeled Artificial Bee Colony clustering algorithm for remotely sensed multi-band image segmentation is explored and evaluated. Data sets of interest include remotely sensed multi-band RGB imagery, which subsequent to classification is analyzed and assessed for accuracy. Results demonstrate that the algorithm exhibits improved accuracy, when compared to its classical counterpart. Moreover, solutions are enhanced via introduction of the quantum state machine, which provides random initial food sources and variables as input to the Artificial Bee Colony algorithm, and quantum operators, which bring about convergence and maximize local search space exploration. Typically, the algorithm has shown to produce better solutions.

A Quantum-Modeled Fuzzy C-Means clustering algorithm for remotely sensed multi-band image segmentation is explored and evaluated. Data sets of interest include remotely sensed multi-band imagery, which subsequent to classification is analyzed and assessed for accuracy. Results demonstrate that the algorithm exhibits improved accuracy, when compared to its classical counterpart. Moreover, in general, the solution is enhanced via introduction of the quantum state machine in and of itself, which provides random fuzzy membership input to the Fuzzy C-Means soft partitioning algorithm, while the addition of quantum operators provide additional contributions to solution diversity. Typically, when evaluated for cluster validity, the algorithm has shown to produce effective solutions.

Recent research shows that the engineering students have problems connecting the required computation to a conceptual understanding, as well as translating a graphical understanding of the process to a symbolic mathematical representation, especially when handling the multiple steps of the procedures. The students are usually able to perform sequences of the underlying calculations but cannot piece together the higher conceptual relationship that drives these procedures. This work-in-progress paper presents a viable approach and a new teaching and learning paradigm to enhance the effectiveness of teaching fast Fourier transform and significantly improve the learning outcomes. By integrating the mobile and cloud computing technologies, we are developing a handheld real-world relevance laboratory that includes an integrated learning module for Fourier transform and a shared intelligent project repository to host the module and real-world relevant data. This development is expected to overcome the intellectual inaccessibility of transform techniques and tackle the challenges in existing approaches: the prohibitive cost of project-based approach; limited access and real-world relevance data in simulation-based approach, and unreliable and unsustainable support.

Secure communication has traditionally been ensured with data encryption, which has become easier to break than before due to the advancement of computing power. For this reason, information hiding techniques have emerged as an alternative to achieve secure communication. In this research, a novel information hiding methodology is proposed to deliver secure information with the transmission/broadcasting of digital video. Secure data will be embedded within the video frames through vector quantization. At the receiver end, the embedded information can be extracted without the presence of the original video contents. In this system, the major performance goals include visual transparency, high bitrate, and robustness to lossy compression. Based on the proposed methodology, the authors have developed a novel synchronization scheme, which ensures audio/video synchronization through speech-in-video techniques. Compared to existing algorithms, the main contributions of the proposed methodology are: (1) it achieves both high bitrate and robustness against lossy compression; (2) it has investigated impact of embedded information to the performance of video compression, which has not been addressed in previous research. The proposed algorithm is very useful in practical applications such as secure communication, captioning, speech-in-video, video-in-video, etc.

In remote sensing researches, the curse of dimensionality is one greatly difficult classification problem. Many studies have demonstrated that multiple classifier systems, such as the random subspace method (RSM), can alleviate small sample size and high dimensionality concern and obtain more outstanding and robust results than a single classifier on extensive pattern recognition issues. A dynamic subspace method (DSM) was proposed for constructing component classifiers with adaptive subspaces to adjust the shortcomings of RSM based on re substitution accuracy by applying each classifier. However, the performances of SVMs are based on choosing the proper kernel functions or proper parameters of a kernel function. The objective of this research is to develop a novel ensemble technique based on support vector machines (SVMs) via the optimal kernel method, and propose a novel subspace selection mechanism, named the kernel-based dynamic subspace method (KDSM), to improve DSM on automatically determining dimensionality and selecting component dimensions for diverse subspaces. Experimental results show a sound performance of classification on the famous hyperspectral images, Washington DC Mall.

In this study, a novel spatial information based support vector machine for hyperspectral image classification, named spatial-contextual semi-supervised support vector machine (SC3SVM), is proposed. This approach modifies the SVM algorithm by using the spectral information and spatial-contextual information. The concept of SC3SVM is to utilize other information, obtain from the pixels of a neighborhood system in the spatial domain, to modify the effective of each patterns. Experimental results show a sound performance of classification on the famous hyperspectral images, Indian Pine site. Especially, the overall classification accuracy of whole hyperspectral image (Indian Pine site with 16 classes) is up to 96.4%, the kappa accuracy is up to 95.9%.

Combining multiple classifiers is a natural way to discover useful information and improve the performances of individual classifiers. It's based on the combination of the outputs of an ensemble of different classifiers. When interactions exist in combining multiple classifiers, fuzzy integral would be a valid method to fuse multiple classifiers. In this fuzzy fusion approach, the fuzzy measure plays an important role. Liu proposed a novel fuzzy measure, L-measure, which is more sensitive than some common measures, like ¿-measure, P-measure and V-measure. In this paper, we would combine the multiple classifiers by Choquet integral wtih this Z-measure.

This paper introduces a new adaptive fuzzy clustering algorithm which combines the capability of fuzzy mathematics and adaptation. This adaptive capability is achieved by using the mechanism of splitting and merging. Unlike most of the fuzzy clustering algorithms which require a priori knowledge about the number of classes in the dataset, this new algorithm can learn the number of classes dynamically. It also gives the higher accuracy of clustering results with fuzzy mathematics. A comparison with the K-Means, ISODATA, Fuzzy C-Means and Possibilistic C-Means shows that the algorithm is effective in image segmentation. The algorithm also enhances the adaptive capability of the ISODATA.

Many studies have demonstrated that multiple classifier systems, such as random subspace method, obtain more outstanding and robust results than a single classifier. In this study, we propose a novel RSM framework which is composed of two parts. The first part is the construction of a weighted RSM, where weights are given by two classifier-based distributions. One is the feature weighting distribution, and the other is the subspace dimensionality distribution that helps for dynamically selecting the size of subspace with respect to the employed classifiers. The second part is to introduce the spatial information estimated by the Markov random filed theory into the Bayesian classifiers used in the framework. The real data experimental results show that the proposed framework obtains satisfactory performances, and the classification maps remarkably produce fewer speckles.

Usually feature extraction is applied for dimension reduction in hyperspectral data classification problems. Many studies show that nonparametric weighted feature extraction (NWFE) is a powerful tool for extracting hyperspectral image features. The detection of class boundaries is an important part in NWFE and the weighted mean was defined for this purpose. In this paper, a kernel-based feature extraction is proposed based on a new class boundary detection mechanism. The soft-margin support vector machine (SVM) binary classifier and the support vector domain description (SVDD) are applied to detect the boundaries between two classes and one class, respectively. The results of real data experiments show that the proposed method outperforms original NWFE.

This paper explores the effectiveness of the rough set theory in multispectral image classification. A new multispectral image classification approach is proposed based on the rough set theory which uses upper and lower bounds for the class description. Rough set theory is used for classification rules extraction. A comparison of this method with the parallelepiped classifier, where the former uses the concept of cuts and the later uses the maximum and minimum values, is compared. Preliminary experimental results show that the proposed classifier is effective for multispectral image classification.

Fuzzy clustering is an approach using the fuzzy set theory as a tool for data grouping, which has advantages over traditional clustering in many applications. Many fuzzy clustering algorithms have been developed in the literature including fuzzy c-means and possibilistic clustering algorithms, which are all objective-function based methods. Different from the existing fuzzy clustering approaches, in this paper, a general approach of fuzzy clustering is initiated from a new point of view, in which the memberships are estimated directly according to the data information using the fuzzy set theory, and the cluster centers are updated via a performance index. This new method is then used to develop a generalized approach of possibilistic clustering to obtain an infinite family of generalized possibilistic clustering algorithms. We also point out that the existing possibilistic clustering algorithms are members of this family. Following that, some specific possibilistic clustering algorithms in the new family are demonstrated by real data experiments, and the results show that these new proposed algorithms are efficient for clustering and easy for computer implementation.

In this study, the performances of hyperspectral image classification using three max-cut based BHC schemes are compared. The first one is original BHC scheme and the second is applied feature extraction in the root node. The last is applied feature extraction in the each non-leaf node. The real data experimental results show that applying feature extraction in the each non-leaf node is the best strategy. Among the combinations of two feature extraction methods, DAFE and NWFE and three base classifiers, using NWFE and ML base classifier can reach the best performance.

Feature extraction is often applied for dimensionality reduction in hyperspectral data classification problems to mitigate the Hughes phenomenon. Some studies had proven that nonparametric weighted feature extraction (NWFE) is a powerful tool to extract well-described features for classification. NWFE concentrates only on the separability of spectral data, however, in many remotely sensed images, objects on the ground are much greater than one pixel. Hence, neighboring pixels are more likely to belong to the same class and form a homogeneous region. We present a scheme to fuse spatial information into NWFE, and from the real data experiments, we can find the proposed method outperforms the original NWFE.

In this paper, semi-supervised classification concept is applied to Discriminant Analysis Feature Extraction (DAFE) and Nonparametric Weighted Feature Extraction (NWFE). The proposed semi-supervised DAFE and NWFE use the information of both labeled and unlabeled data in an iterative process. Experimental results of hyperspectral data show that the proposed feature extraction methods can improve the classification performance significantly with limited training samples.

The parallelepiped classifier is one of the widely used supervised classification algorithms for multispectral images. The threshold of each spectral (class) signature is defined in the training data, which is to determine whether a given pixel within the class or not. To avoid involving the analyst for the training data selection, this paper is to study whether the threshold of parallelepiped classifier can be automatically determined by using natural evolution process - genetic algorithms (GAs). In other words, our goal is to create an unsupervised multispectral parallelepiped classifier with the help of genetic algorithms. In this algorithm, we also use a new approach to estimate the initial range. Preliminary experimental results with different parameters for genetic algorithms and a comparison with the supervised parallelepiped classifier are provided.

We compared three image classifiers which incorporate contextual information to classify each pixel in the raw images in this study. These procedures incorporate contextual information by using different features and classify the pixel into one of several predefined classes based on these features. These spatial classifiers strive to capture the spatial relationships encoded in the aerial photograph. The determination of the window size is a challenging issue in these spatial classifiers. Preliminary experimental results are provided in this report.

A different variety of services requiring different levels of quality of service (QoS) need be addressed for mobile users of the next generation wireless system (NGWS). An efficient handoff technique with intelligent admission control can accomplish this aim. In this paper, a new, intelligent handoff scheme using distributed genetic algorithms (DGA) is proposed for NGWS. This scheme uses DGA to achieve high network utilization, minimum cost and handoff latency. A performance analysis is provided to assess the efficiency of the proposed DGA scheme. Simulation results show a significant improvement in handoff latencies and costs over traditional genetic algorithms and other admission control schemes.

We developed a spatial image classifier which incorporates contextual information to classify each pixel in the raw images. The procedure incorporates contextual information by using the average of the pixels from a peer group of each pixel and classifying this pixel into one of several predefined classes based on the average. The spatial classifier strives to capture the spatial relationships encoded in the aerial photograph. The Fisher discriminant was used to select the peer group, that means the determination of the window size and number of pixels for each pixel being classified in this spatial classifier. Experimental results are provided.

The paper outlines two image enhancement operators for gray scale and color images. A modified cosine function was developed for image enhancement, in which some enhanced images may appear a little darker if the average pixel value in the image generally falls below the intermediate value (which is 128 in most images used in the experiments (256 gray-level)). A Semi-Histogram-Equalization (SHE) method is then proposed for enhancement so that regardless of how dark or bright the image is, it would give a good contrast enhancement. Experimental results show that the proposed operators can enhance gray scale and color images effectively.

This paper describes a genetic differential competitive learning algorithm, which is proposed to prevent fixation to the local minima and improve the unsupervised training results for the classification of remotely sensed data. The differential competitive learning (DCL) combines competitive and differential-Hebbian learning and represents a neural version of adaptive delta modulation. This learning law uses the neural signal velocity as a local unsupervised reinforcement mechanism. The Jeffries-Matusita (J-M) distance, which is a measure of statistical separability of pairs of the 'trained' clusters, is used for the evaluation of the proposed algorithm. The Landsat Thematic Mapper (TM) data will be used for simulation to show the effectiveness of the algorithm.