Search

This paper is dedicated to developing high-efficiency face detection and tracking method for big dynamic crowds or numerous pedestrians. Three modules constitute the proposed method, i.e., face candidate generation, face candidate verification, and face target tracking. In this work, face candidates are localized using the features of the face area, edge information, and skin color. Non-face parts in the face candidates are further verified by the C-SVM learning model and then removed, by which the face targets can be generated with lower computation-complexity and satisfactory accuracy than other approaches. Finally, the face targets are tracked by an efficient and reliable searching scheme for improving the effective face detection rate. Experimental results show that the average face detection rate (FDR) of 85%, average effective FDR of 95%, a frame rate of 28–66 frames per second (fps), and about 30 faces detected per frame are obtained from various test videos with big dynamic crowds or numerous pedestrians, indicating the feasibility of the proposed method to achieve unconstrained face detection with high-efficiency and cost-effectiveness. This result makes the proposed method more attractive for the video surveillance system as compared to other approaches, especially in the high computational complexity-based methods.

The complexity of enterprises’ annual reports has increased significantly under supervision. Although the interactive relations of enterprises attracted interest both among scholars and practitioners, however, the decision support information is scattered in heterogeneous and enormous data. Here introduces a new learnable deep network—which implements an inference for entity- and relation-centric representations of multilayer, dynamical but sparse system. The results show the approach supporting accurate predictions from real data extracted from financial reports. Indeed, the method shows robustness with incomplete data, thus representing a valuable tool for gaining insights on privacy-protected financial systems.

The three-parameter Logistic model (3PLM) and the four-parameter Logistic model (4PLM) have been proposed to reduce biases in cases of response disturbances, including random guessing and carelessness. However, they could also influence the examinees who do not guess or make careless errors. This paper proposes a new approach to solve this problem, which is a robust estimation based on the 4PLM (4PLM-Robust), involving a critical-probability guessing parameter and a carelessness parameter. This approach is compared with the 2PLM-MLE(two-parameter Logistic model and a maximum likelihood estimator), the 3PLM-MLE, the 4PLM-MLE, the Biweight estimation and the Huber estimation in terms of bias using an example and three simulation studies. The results show that the 4PLM-Robust is an effective method for robust estimation, and its calculation is simpler than the Biweight estimation and the Huber estimation.

In the motor-driven high-voltage circuit breaker, as the power executive part of the operating mechanism, the performance of motor plays a decisive role in the opening and closing speed of the high-voltage circuit breaker. Therefore, the design of a high-power permanent magnet synchronous motor (PMSM) with high torque density and low torque ripple is the basis for the stable operation of high-voltage circuit breakers. According to the current use of PMSM, it is found that the cogging torque caused by the interaction between the permanent magnet field and the cogging of stator, it will cause oscillation at the end position of circuit breaker. According to the practical problems, the method of rotor slanting pole and opening hole is adopted. Through electromagnetic simulation calculation and actual product test, finally, we can get the optimal solution of the cogging design of PMSM for high-voltage circuit breaker.

There is a lot of unstructured text information in the notes to financial reports, also, text mining and information extraction are the difficulties in automatic processing of enterprise financial reports. In order to solve the problem of information extraction and text analysis for cross-paragraph events, this paper use the method of graph theory to construct the whole text features, and form a semantic block perception and classification method that embodies the chapter-semantic block-sentence-vocabulary feature words and their interrelations. Thus, it provides a systematic solution for extracting text information from financial report in large quantities, and offers a practical application scenario.

This paper proposes a model to extract textual information from financial reports automatically. It takes event extraction as the core, maps narrative information of financial reports into the concepts of financial accounting field, and forms the integration of heterogeneous data of distributed financial information. This study shows that the model can identify text events in large-scale financial reporting corpus, automatically extract events and theirs attribute information, and convert them into structured data. Moreover, this paper presents and evaluates the effect of the model in extracting information from annual reports of listed companies. The experimental results turn out that the model provides semantic mapping between text events and domain knowledge concepts, which is reasonable and reliable to be applied in the field of financial statement analysis.

The number of companies using social media to interact with users has been increasing. By analyzing the content generated by users online and mining the information contained therein, companies have applied them to product management and innovation, improving the competitiveness of enterprises. In the face of growing user-generated content, how to achieve effective information mining and transform into product knowledge has posed a challenge to the enterprise. In the background of smart phone products, this paper puts forward the integrated method and key technology of innovation knowledge management based on user-generated content, which should be able to provide a systematic solution for the interactive innovation knowledge management.

This paper presents a driver assistance system for vehicle detection and inter-vehicle distance estimation using a single-lens video camera on urban/suburb roads. The task of vehicle detection on urban/suburb roads is more challenging due to their high scene complexity. In this work, the still area of frame inside the host vehicle is first removed using temporal differencing, followed by detecting vanishing point. Segmentation of road regions is then conducted using vanishing point and road’s edge lines. Shadow regions at the bottoms of vehicles verified using the HOG feature and an SVM classifier are utilized to detect vehicle positions. The distances between the host and its front vehicles are estimated based on the locations of detected vehicles and vanishing point. Experimental results show varied performance of vehicle detection with different scenes of urban/suburb roads and the detection rate can achieve up to 94.08%, indicating the feasibility of the proposed method.

In order to improve the working characteristics of high-voltage circuit breakers and achieve energy saving and emission reduction, a new operating mechanism of 126kV high-voltage vacuum circuit breaker directly driven by three-phase permanent magnet synchronous motor (PMSM) was studied. The simulation model of the control system was established by Matlab/Simulink, and the control system of the direct drive operation mechanism of the vacuum circuit breaker was designed and implemented based on the TMS320F28335 hardware platform. The simulation calculation and experimental results show that the 126kV high-voltage vacuum circuit breaker operating mechanism directly driven by PMSM can realize the opening and closing operation of the circuit breaker according to the preset contact motion curve. The research shows the feasibility of using a motor to drive the 126kV high-voltage vacuum circuit breaker operating mechanism, and can meet the requirements of high-voltage circuit breaker operation.

This paper proposes a robust method for detecting copy-move forgery in images under various JPEG compression and Gaussian noise and blurring attacks. The method comprises feature extraction, feature matching, and duplicate block identification. The fast Fourier transform (FFT), singular value decomposition (SVD), and principal component analysis (PCA) are utilized for feature extraction. Then, the cascading matchers of FFT, SVD, and PCA are adopted for feature matching. Matched blocks are identified using cascade filtering with city block, horizontal, vertical, and frequency filters. Finally, the pixels on upper left corners of detected duplicate blocks are output for visual inspection. The major contributions of this paper are: (1) the proposed method is fully threshold-free; (2) only one-dimensional features are generated using FFT and SVD for feature matching; and (3) a high accuracy rate (> 97 %) is obtained even if the JPEG quality factor Q is 20. In extensive experiments, a 98 % accuracy rate and a 6 % false negative rate were obtained for the worst case of Q?=?20 and a region size of 32?×?32 pixels compared to existing works, indicating the feasibility of the proposed method.

This paper presents a real-time face detection system using a moving camera. The proposed system consists of three modules, including (1) detection of face candidates: Face candidates are generated using the information of skin color, edges, and face area, (2) verification of face candidates: HOG (Histogram of Oriented Gradient) features are generated from face candidates and a two-class C-SVM (Support Vector Machine) classifier with pretrained face samples is employed to determine whether face candidates are real faces or not, (3) face tracking: Overlapping area of two face targets in current and previous frames is estimated to determine whether the tracking will be continuous or not. By use of estimation of face size, the proposed method can avoid a huge amount of computation time that is required by a point-by-point scanning way in conventional methods. Moreover, the accuracy of the face detection can be improved greatly. The proposed system can successfully detect most faces of the crowds in open space, which is beneficial for quickly searching the specified persons to prevent the occurrence of possible criminal events.

Proposed method has good performance for a moving camera without additional sensors.Proposed method works well for tracking overlapping objects with scale changes.Proposed method outperforms the-state-of-art methods. This paper presents an effective method for the detection and tracking of multiple moving objects from a video sequence captured by a moving camera without additional sensors. Moving object detection is relatively difficult for video captured by a moving camera, since camera motion and object motion are mixed. In the proposed method, the feature points in the frames are found and then classified as belonging to foreground or background features. Next, moving object regions are obtained using an integration scheme based on foreground feature points and foreground regions, which are obtained using an image difference scheme. Then, a compensation scheme based on the motion history of the continuous motion contours obtained from three consecutive frames is applied to increase the regions of moving objects. Moving objects are detected using a refinement scheme and a minimum bounding box. Finally, moving object tracking is achieved using a Kalman filter based on the center of gravity of a moving object region in the minimum bounding box. Experimental results show that the proposed method has good performance.

This paper proposes an effective image forgery detection scheme that identifies a tampered foreground or background image using image watermarking and alpha mattes. In the proposed method, component-hue-difference-based spectral matting is used to obtain the foreground and background images based on the obtained alpha matte. Next, image watermarking based on the discrete wavelet transform, discrete cosine transform, and singular value decomposition is used to embed two different watermarks into the foreground and background images, respectively. Finally, the difference between the obtained singular values is used to detect tampering of foreground or background image. Experimental results show that the proposed method performs well in terms of image forgery detection.

Camera tampering and abnormalities are examined for video surveillance system.Brightness, edge details, and histogram information are computationally efficient.The system runs at 20-30frames/s, meeting the requirement of real-time operation.An average of 4.4% of missed events indicates the feasibility of proposed method. Camera tampering may indicate that a criminal act is occurring. Common examples of camera tampering are turning the camera lens to point to a different direction (i.e., camera motion) and covering the lens by opaque objects or with paint (i.e., camera occlusion). Moreover, various abnormalities such as screen shaking, fogging, defocus, color cast, and screen flickering can strongly deteriorate the performance of a video surveillance system. This study proposes an automated method for rapidly detecting camera tampering and various abnormalities for a video surveillance system. The proposed method is based on the analyses of brightness, edge details, histogram distribution, and high-frequency information, making it computationally efficient. The proposed system runs at a frame rate of 20-30frames/s, meeting the requirement of real-time operation. Experimental results show the superiority of the proposed method with an average of 4.4% of missed events compared to existing works.

In recent years, with the great success of sparse representation on face recognition, sparse representation based classification (SRC) and collaborative representation based classification (CRC) are two most popular methods, among which SRC with L1-norm minimization has a relatively better recognition rate but it is time consuming while CRC with L2-norm minimization has a relatively better time efficiency but recognition performance is lower than that of SRC, especially when face images are impaired or undesirable outliers are involved. To achieve both time efficiency as CRC and better recognition performance as SRC, a novel method called adaptable dense representation with CRC (ADR+CRC) is presented, which utilizes ADR for dictionary decomposition using the technique of low-rank matrices recovery to improve the deficiency of CRC that has poor recognition performance when face images are corrupted due to disguise and occlusion. Experimental results show the superiority of the proposed method when compared with the state-of-the-art methods.

An efficient method for detecting moving vehicles based on the filtering of swinging trees and raindrops is proposed. To extract moving objects from the background, an adaptive background subtraction scheme with a shadow elimination model is used. Swinging trees are removed from foreground objects to reduce the computational complexity of subsequent tracking. Raindrops are removed from foreground objects when necessary. Performance evaluations are carried out using seven real-world traffic image sequences. Experimental results show average recognition rates of 96.83% and 97.20% for swinging trees and raindrops, respectively, indicating the feasibility of the proposed method.

A scheme based on a difference scheme using object structures and color analysis is proposed for video object segmentation in rainy situations. Since shadows and color reflections on the wet ground pose problems for conventional video object segmentation, the proposed method combines the background construction-based video object segmentation and the foreground extraction-based video object segmentation where pixels in both the foreground and background from a video sequence are separated using histogram-based change detection from which the background can be constructed and detection of the initial moving object masks based on a frame difference mask and a background subtraction mask can be further used to obtain coarse object regions. Shadow regions and color-reflection regions on the wet ground are removed from the initial moving object masks via a diamond window mask and color analysis of the moving object. Finally, the boundary of the moving object is refined using connected component labeling and morphological operations. Experimental results show that the proposed method performs well for video object segmentation in rainy situations.

Kernel-level Rootkit injects malicious code in the system kernel and realizes vicious function by modifying kernel code and data. The emergence of virtualization technology makes Rootkit hide itself and accesses system resources effectively. This paper proposes a kind of Rootkit based on VMM (VMRootkit) which can hide system process and system cannot find it. For realizing it, we firstly puts forward formal model of VMRootkit about cooperative concealment combined with Trojans thought about cooperative concealment, then researches the process switch and data structure of operating system and realized prototype of it, which can monitor system process and realize process depth hiding by modifying data structure related process view. Experiment shows that VMRootkit prototype meets the idea of cooperative concealment, can realize process depth hiding and has better hidden property than kernel-level Rootkit.

This paper is dedicated to the front vehicle detection and its distance estimation using one single-lens car video camera on the urban & suburb road roads. The proposed method mainly consists of road area detection and the front vehicle detection & distance calculation. Firstly, Hough transform is used to detect lines in which the appropriate straight lines are selected and their intersection points are exploited for obtaining the vanishing point. Then, both strong right and left edges are extracted and connected to the vanishing point for segmenting the road area, and on such area the bottom shadow of the front vehicle is utilized to locate the vehicle's position. Finally, the distance between the host vehicle and the front vehicle is calculated based on the position of a vehicle and the vanishing point. Experimental results show that the proposed technique can moderately detect the front vehicles on the urban & suburb roads with detection rate of 78% at least.

This paper presents a visual surveillance scheme for cage aquaculture that automatically detects and tracks ships (intruders). For ship detection and tracking, we propose a robust foreground detection and background updating to effectively reduce the influence of sea waves. Furthermore, we propose a fast 4-connected component labeling method to greatly reduce the computational cost associated with the conventional method. Wave ripples are removed from regions with ships. An improved full search algorithm based on adaptive template block matching with a wave ripple removal is presented to quickly, accurately, and reliably track overlapping ships whose scales change. Experimental results demonstrate that the proposed schemes have outstanding performance in ship detection and tracking. The proposed visual surveillance system for cage aquaculture triggers an alarm if intruders are detected. The security of cage aquaculture can be increased. The proposed visual surveillance can thus greatly help the popularization of cage aquaculture for ocean farming.

Research highlights? We proposed a method for hand gesture recognition based on Gabor filters and support vector machines under varying illumination. ? The proposed method is robust against varying illumination and insensitive to hand-pose variations. ? The proposed method allows users to wear either a long- or short-sleeve shirt. ? The recognition rates are improved from 72.8% to 93.7% by using the proposed hand-pose correction module. In this paper, we present a novel approach for hand gesture recognition based on Gabor filters and support vector machine (SVM) classifiers for environments with varying illumination. The proposed method (1) is robust against varying illumination, which is achieved using an adaptive skin-color model switching method; (2) is insensitive to hand-pose variations, which is achieved using a Gabor filter-based gesture angle estimation and correction method; (3) allows users to wear either a long- or short-sleeve shirt, which is achieved using a method that segments the hand from the forearm. To evaluate the robustness of the proposed method, we created a database of hand gestures in realistic conditions. A recognition rate of 96.1% was achieved using the proposed method. A dynamic gesture recognition system is also presented for real-life conditions. In the proposed system, the recognition results improved from 72.8% to 93.7% when the hand-pose correction module was used, indicating that using the responses of Gabor filters to estimate the hand-pose angle is effective.

Face recognition is quite challenging especially in a varying illumination scenario. In this paper, an effective method for face recognition based on dual-tree complex wavelet transform under low illumination environments is proposed. To alleviate the effect of lighting change on face image, logarithmic transform is performed, and a wavelet denoising model is then used to remove the low-frequency components in the details of LH, HL, and HH subbands. To enhance facial features, six difference images obtained from the four LL subbands are utilized and averaged to a so-called mean face. Results show that serious cast shadows on face images can be effectively removed such that face recognition rate can be greatly improved. To verify the feasibility of the proposed method, Yale face database B is used. Experimental results show that an overall recognition rate of 96% can be achieved.

In a fault-tolerant on-chip network, overhead due to the Automatic Repeat-reQuest (ARQ) operation may significantly impact the performance. In this work, an intelligent prefetching scheme that can effectively mitigate the performance hits from the ARQ overhead is proposed. Specifically, a packet-based retransmission scheme is incorporated to improve payload utilization while using Error-Detecting Code (EDC). A Virtual Cut-Through (VCT) switching is implemented to alleviate drastic performance degradation in dealing with the damaged data caused by noise upsets during packet transmissions. Potential performance advantages in terms of packet latency and transmission throughput were carefully analyzed to validate the practice of our proposed prefetching scheme in the designed network on-chip systems.

The detection of copy-move forgery image is important in the field of blind image forensics because it is of pure image processing technique without any support of embedded security information. The proposed method consists of a boosting scheme, feature extraction and similarity matching for the detection of duplicated regions. The boosting scheme comprises an estimation of dark channel, histogram equalization and grayscale layering, by which the number of image blocks on each subimage layer can be dramatically reduced so that the time efficiency of subsequent lexicographical sorting and similarity matching can be greatly improved. Experimental results show that the proposed boosting scheme can significantly enhance the computation efficiency and have a good detection rate. Moreover, the propose method is robust to any angles rotation attack.

A nighttime driver-assistance system is developed in this study. Through lane detection, neighboring vehicle detection, lane deviation detection, and front vehicle distance estimation of real-time road images retrieved using a charge-coupled device camera, the developed system can issue timely alarms to warn drivers and increase their driving safety at nighttime. In lane detection, the developed system employs clear edge information and lane orientation, and performs edge operations and lane marking using Hough Transform. Vehicle detection is performed by recognizing vehicle characteristics (e.g., tail lights and symmetry) and colors. The front vehicle distance is estimated by counting the number of lane segments. The experimental results of numerous highway image tests show that the developed system can effectively facilitate safe driving in a common environment. Keywords: Driver-assistance system, Lane detection, Neighboring vehicles detection, Lane deviation detection, and Vehicle distance estimation.

In this paper, we propose an innovative approach for face recognition based on collaborative image similarity assessment (CISA). In the proposed method, the test sample is first represented by a linear combination of all the training samples for each face class. The classification task is then conducted using the similarity measures including structure similarity index measure (SSIM), root mean square (RMS), and similarity assessment value (SAV). Since CISA is only one phase, it is computationally efficient when comparing with the method of two-phase test sample sparse representation (TPTSR). To verify the performance of face classification, two popular face databases of the ORL and FERET are used. Results show that CISA is comparable with TPTSR on the classification rates for ORL database. However, CISA greatly outperforms TPTSR on the evaluation of the FERET database. Moreover, only 276.4ms on an average is required for CISA in the classification of each test sample but it needs 800.8ms for TPTSR.

In this paper, we present a novel method for gait recognition of different people groups based on Fourier descriptors (FDs) and support vector machine (SVM). The proposed method involves the procedures of background modeling, extraction of gait silhouettes by background subtraction, shadow removal, representation of gait silhouettes using the FDs, and recognition of human gaits by an SVM Classifier. To evaluate the performance of the proposed method, 5 different people groups with 15 video sequences, including the pregnant, the children, the adult, the people with a walking stick, and the aged, are utilized. Experimental results show that the correct recognition rate (CRR) of 81% is obtained using only the first 20 coefficients of FDs, indicating the feasibility of the proposed method.

The identification of bloodstains in color images is critical in some accident investigation cases. A suspected bloodstain segmentation method was proposed in this study. The algorithm transforms the RGB to the YCgCr color space after the image undergoes color histogram equalization. Determination of the decision thresholds (Cg, Cr) is based on the color distribution of the sample images in Cg-Cr plane. To accelerate bloodstain pixel detection, the transformed thresholds are used (Cgi¦, Cri¦). After the bloodstain pixels are detected the fast 8-connected component labeling method is to formulate the clusters. The algorithm was tested by using the 75 images from accident scenes. The results show that su spected bloodstains can be detected successfully, and thus that this approach can be used as the basis for bloodstain identification in color images.

In time critical system, message scheduling plays an important role to arbitrate fair service among all competing messages, where messages are conditioned on different timing constraints. There already exist many algorithms, including static and dynamic scheduling, to resolve these problems. In this paper, we extend our recent work by presenting a fuzzy inference system (FIS) as a message classifier for scheduling. The FIS with adaptation strategy of parameter adaptation and structure identification can always result in smaller untimely service ratios (USRs) and small number of fuzzy rules, especially when the traffic load is heavy. Moreover, the resulting FIS model would not remain a black box. Instead, the implicit scheduling knowledge can be interpreted in terms of linguistic fuzzy sets. The efficiency of the proposed method is further examined by comparing with several traditional scheduling methods. Simulation results confirm our claims consistently.

In this paper, we propose a method of eye detection based on skin color analysis under varying illumination. The proposed method consists of several phases, including color conversion, skin color segmentation and face mask calculation, facial feature extraction and eye candidate determination, and detection of human eyes. To eliminate the effect of lighting change on the performance of eye detection, color conversion is first performed. Face mask calculation based on skin color segmentation is then carried out to reduce the possible searching region of human eyes. Eye candidates detected using facial features are used to define the possible human eyes. Human eyes are thus detected by the geometric features of gravity and spatial centers for these eye candidates. Results show that the proposed method works well for faces with different poses and multiple faces under varying illumination. The eye detection time of 21.8 ms is achieved for an image of size 213x320 pixels, indicating computational efficiency of the proposed system.

In this paper, we propose a method of face detection based on feature analysis and edge detection against skin color-like backgrounds. The proposed method consists of three phases including image preprocessing, skin color segmentation, and determination of face candidates. First, the objects in foreground are separated by image preprocessing. Second, the non-skin color objects are rejected by the method of skin color segmentation. Finally, both the elimination of skin color-like blobs in backgrounds and face detection are performed by the method of determination of face candidates. Experimental results show that the proposed method can effectively detect most of faces in skin color-like backgrounds. In addition, detection of face images with pose and expression variations or against dark backgrounds are also carried out.

This paper proposes an automatic video object segmentation with opacity estimate to obtain good alpha mattes and the foreground objects. In the proposed method, the trimap is automatically generated from the result obtained using video object segmentation based on motion information without background construction. Next, closed-form matting is used with the automatically generated trimaps to yield the alpha mattes and the foreground objects. Experimental results show that the proposed method has good performance in video object segmentation with opacity estimate.

This paper presents a modified spectral matting to obtain automatic and accurate image matting. Spectral matting is the state-of-the-art image matting and also a milestone in theoretic matting research. However, using spectral matting without user guides, the accuracy is usually low. The proposed modified spectral matting effectively raises the accuracy. In the proposed modified spectral matting, the palette-based component classification is proposed to obtain the reliable foreground and background components. In contrast to the spectral matting, based on these reliable foreground and background components, the accuracy of obtained alpha matte is greatly increased. Experimental results show that the proposed method has better performance than the spectral matting. Therefore, the proposed image matting is very suitable for image and video editing.

In this paper, we propose a novel algorithm based on Gabor wavelets and support vector machine (SVM) for recognition of tire tread patterns. Input tire images are first preprocessed by morphological opening to enhance the features (or textures) on tire surface. The grooves in tire surface are salient important features for a tire matching system. We detect the tire tread patterns of being grooved or wavy and use this feature to train various SVM classifiers. The features of tire tread patterns are then represented by Gabor wavelets, and feature extraction is further carried out by principal component analysis (PCA). Finally, the matching processes are achieved by the classifiers of SVM, Euclidean distance and cosine distance. Result shows that the recognition rate of 60% for tire images can be obtained by the SVM classifier when 15 tire tread patterns are used.

This paper proposes a three-stage scheme to obtain real-time and reliable face detection in intersecting monitoring. The proposed three-stage scheme is mainly based on skin color and facial features. In the first stage scheme, skin-color model is used to obtain skin regions. The second stage scheme uses face template measure to obtain face candidates. Finally, facial features are measured to detect faces from face candidates. Experimental results show that the proposed method has good performance in face detection.

In this paper, an adaptive skin color model switching based on AdaBoost method for face tracking is proposed. Possible skin clusters under illumination varying scenes are detected by an optimal skin color model, which is adaptively selected by a well-defined quality measure. The possible facial candidates are further validated by AdaBoost to determine whether human faces exist in video sequences or not. The tracking sequences reveal that good and robust results are obtained from dim- to profile- to back-light scenarios. The performance of the proposed method can achieve an average tracking time of about 145.4 ms/frame and a detection rate of 94.4%.

In this paper we present a novel method for hand gesture recognition based on Gabor filters and support vector machine (SVM). Gabor filters are first convolved with images to acquire desirable hand gesture features. The principal components analysis (PCA) method is then used to reduce the dimensionality of the feature space. With the reduced Gabor features, SVM is trained and exploited to perform the hand gesture recognition tasks. To confirm the robustness of the proposed method, a dataset with large posed-angle (≫45 deg.) of hand gestures is created. The experiment result shows that the recognition rate of 95.2% can be achieved when SVM is used. A real-time video system for hand gesture recognition is also presented with a processing rate of 0.2 s for every frame. This result proves the efficiency and superiority of the proposed Gabor-SVM method.

This paper proposes an effective method for video object segmentation without background construction. In the proposed method, the coarse foreground extraction and fine foreground extraction are obtained using the motion information, edge information, and gradient-variation information which are first evaluated by two successive frames. Next, the video object is extracted using the horizontal/vertical filling scheme based on the coarse foreground extraction and fine foreground extraction. Finally, video object refining is used to obtain the more accurate video object. Experimental results show that the proposed method has good performance in sensitivity, specificity, and spatial accuracy.

Visualization is the process of transforming data, information, and knowledge into visual form, making use of humans' natural visual capabilities. In the area of geological data visualization, the real-time dynamic analysis is very important for the users. Real-time data visualization and real-time feature magnification are two key points in real-time dynamic analysis. The relative study on visualization and feature magnification in this area is still rare. In this paper we design a data visualization system for real-time geological information visualization. Firstly, this paper provides two kinds of feature magnification methods: overview plus detail and focus plus context, which can present specific detailed information for users' observation and analysis. In addition, this paper discusses how to meet the real-time requirement for our data visualization system.

The interaction of product or service attributes may influence the preferences of consumers. If the numbers of stimuli (profiles) are large, then the combinations of the levels present difficulty for consumers to rank order in terms of their preference. In this chapter, our purpose is to propose a ranking and selection procedure to select the most preferred attributes in such situations. We use the orthogonal array design to build a suitable MANOVA model in conjoint analysis. A split-plot experiment is conducted and the method of moments is used to obtain unbiased estimators of the parameters. The fitness of the MANOVA models is also studied. Some examples based on a preference study of a household electrical appliance are discussed.

This paper proposes a method for recognizing human identity using gait features based on Gabor wavelets and modified gait energy images (GEIs). Identity recognition by gait generally involves gait representation, extraction, and classification. In this work, a modified GEI convolved with an ensemble of Gabor wavelets is proposed as a gait feature. Principal component analysis is then used to project the Gabor-wavelet-based gait features into a lower-dimension feature space for subsequent classification. Finally, support vector machine classifiers based on a radial basis function kernel are trained and utilized to recognize human identity. The major contributions of this paper are as follows: (1) the consideration of the shadow effect to yield a more complete segmentation of gait silhouettes; (2) the utilization of motion estimation to track people when walkers overlap; and (3) the derivation of modified GEIs to extract more useful gait information. Extensive performance evaluation shows a great improvement of recognition accuracy due to the use of shadow removal, motion estimation, and gait representation using the modified GEIs and Gabor wavelets.

This paper proposes an automatic video matting method that uses hybrid video object segmentation and closed-form matting. Hybrid video object segmentation, based on background construction-based video object segmentation and foreground extraction-based video object segmentation, is used to obtain reliable foreground objects. Next, the trimaps are automatically generated using two different scanning windows from the hybrid video object segmentation results. Finally, closed-form matting is used, with the automatically generated trimaps, to yield the alpha mattes and the foreground objects with an opacity estimate. Experimental results show that the proposed method outperforms state-of-the-art automatic video matting methods based on closed-form matting.

This study uses the Fuzzy Multilayered Analytic Hierarchy Process (FMAHP) method with group decision-making procedures to seek the consensus of experts from different backgrounds, as an objective and reasonable public resource disposition strategy. FMAHP is a method that combines fuzzy theory and multi-layer hierarchy process. Furthermore, complete ranking method and grey association analysis are employed in the group decision-making perspective, for the purpose of evaluating the consistency of these results from those of FMAHP. The results show that the FMAHP method achieves the degree of reliability in government expenditure decision-making.

In Section 5.2, we deal with two important aspects of multiple decision rules for the robustness of rules to determine the optimal sample size. The most economic rule, roughly defined to be a rule with least sample size to meet the goal, is described later. Robustness is, of course, a very important consideration in practical applications.

In this chapter, we shall study multiple decision procedures in terms of hypotheses-testing problems. First, we discuss the conditional confidence approach of Kiefer which can be used to improve Neyman-Pearson (NP) formulation. This is done in Section 6.2. Using this approach, we describe conditional selection procedures and their relation with classical selection rules. Later, we discuss the theory of multiple comparisons for some appropriate alternative hypotheses. In Section 6.3, we consider an optimal criterion to improve the power of the individual test. Using this approach, we derive selection rules based on tests. Multiple range tests are studied in Section 6.4. A discussion of the multistage comparison procedures is provided in Section 6.5.

In Section 4.2, we define and study invariant procedures. We describe optimal rules in the class of invariant procedures. Various optimality criteria of the goodness of a rule are mentioned. Various selection procedures based on sample values and their ranks are also studied. Different types of loss functions for subset selection approach for the derivation of selection rules are described and studied. Optimality of classical selection rules in terms of their closeness to the Bayes rules is considered.

In this chapter, we shall discuss the modified principles of minimax criteria (such as г-minimax and restricted minimax) for multiple decision problems.

In the theory and practice of statistical inference, multiple decision problems are encountered in many experimental situations. The classical methods for analyzing data do customarily employ hypothesis testing in most situations. In such cases, when the hypothesis is rejected, one wants to know in which of a number of possible ways the actual situation (true state of nature) differs from the one postulated by the null hypothesis. If, in the formulation of the problem, we consider only two decisions (reject or not reject the hypothesis), we will not only neglect to differentiate between certain alternative decisions but may also be using an inappropriate acceptance region for the hypothesis. Moreover, the traditional approach to hypotheses testing problems is not formulated in a way to answer the experimenter’s question, namely, how to identify the “best” (in some sense) population? For example, the method of the least significant differences based on the t-test has been used in the past to detect differences between the true unknown means of different varieties and thereby choosing the population which is the “best”, say, the one with the largest mean. But this method is indirect, less efficient and does not provide an overall probability of a correct decision. The remark (criticism) is also valid, to some extent, for methods based on multiple comparison techniques. The traditional approach does not allow for a decision if the null hypothesis is not rejected.