In recent years, local pattern based features have attracted increasing interest in object detection and recognition systems. Local Binary Pattern (LBP) feature is widely used in texture classification and face detection. But the original definition of LBP is not suitable for human detection. In this paper, we propose a novel feature named gradient local binary patterns (GLBP) for human detection. In this feature, original 256 local binary patterns are reduced to 56 patterns. These 56 patterns named uniform patterns are used for generating a 56-bin histogram. And gradient value of each pixel is set as the weight which is always same in LBP based features in histogram calculation to computing the values in 56 bins for histogram. Experiments are performed on INRIA dataset, which shows the proposal GLBP feature is discriminative than histogram of orientated gradient (HOG), Semantic Local Binary Patterns (S-LBP) and histogram of template (HOT). In our experiments, the window size is fixed. That means the performance can be improved by boosting methods. And the computation of GLBP feature is parallel, which make it easy for hardware acceleration. These factors make GLBP feature possible for real-time pedestrian detection.

In this paper, we propose a high accuracy motion estimation method based on the spatio-temporal gradient method using high frame-rate images. In the method, we adopt spatial gradients with low estimated errors by the previous motion vectors. In addition, we evaluate the proposed method and confirm the effectiveness. Finally, we apply the method to super-resolution as an application of the proposed method.

This paper presents a new hybrid speed function needed to perform image segmentation within the level-set framework. The proposed speed function uses both the boundary and region information of objects to achieve robust and accurate segmentation results. This speed function provides a general form that incorporates the robust alignment term as a part of the driving force for the proper edge direction of an active contour, an active region term derived from the region partition scheme, and the smoothing term for regularization. First, we use an external force for active contours as the Gradient Vector Flow field. This is computed as the diffusion of gradient vectors of a gray level edge map derived from an image. Second, we partition the image domain by progressively fitting statistical models to the intensity of each region. Here we adopt two Gaussian distributions to model the intensity distribution of the inside and outside of the evolving curve partitioning the image domain. Third, we use the active contour model that has the computation of geodesics or minimal distance curves, which allows stable boundary detection when the model's gradients suffer from large variations including gaps or noise. Finally, we test the accuracy and robustness of the proposed method for various medical images. Experimental results show that our method can properly segment low contrast, complex images.

This work presents a real-time human detection system for VGA (Video Graphics Array, 640480) video, which well suits visual surveillance applications. To achieve high running speed and accuracy, firstly we design multiple fast scalar feature types on the gradient channels, and experimentally identify that NOGCF (Normalized Oriented Gradient Channel Feature) has better performance with Gentle AdaBoost in cascaded classifiers. A confidence measure for cascaded classifiers is developed and utilized in the subsequent tracking stage. Secondly, we propose to use speedup techniques including a detector pyramid for multi-scale detection and channel compression for integral channel calculation respectively. Thirdly, by integrating the detector's discrete detected humans and continuous detection confidence map, we employ a two-layer tracking by detection algorithm for further speedup and accuracy improvement. Compared with other methods, experiments show the system is significantly faster with 20 fps running speed in VGA video and has better accuracy as well.

This paper proposes a method of accurately detecting the boundary of narrow stripes, such as lane markings, by employing gradient cues of edge points. Using gradient direction cues, the edge points at the two sides of the boundary of stripes are classified into two groups before the Hough transform is applied to extract the boundary lines. The experiments show that the proposed method improves significantly the performance in terms of the accuracy of boundary detection of narrow stripes over the conventional approaches without edge point grouping.

This report focuses on a design method for gradient index (GRIN) lens antennas with controllable aperture field distributions. First, we derive differential equations representing optical paths in a gradient index medium with two optical surfaces by using geometrical optics, and then we formulate a novel design method for GRIN lens antennas based on these equations. The Levenberg-Marquardt algorithm is applied as a nonlinear least squares method to satisfy two conditions-focusing and shaping the aperture field distribution-thus realizing a prescribed radiation pattern. The conditions can be fulfilled by optimizing only the index (or permittivity) distribution, whereas the shapes of the optical surfaces remain as free parameters that can be utilized for other purposes, such as reducing reflection losses that occur on the surfaces, as illustrated in this report. A plano-concave GRIN lens is designed as an example, applying the proposed method, to realize a sidelobe level of -30 dB pseudo Taylor distribution, and a maximum sidelobe level of -29.1 dB was observed, indicating it is sufficiently accurate for practical use. In addition, we discuss the convergence of this method considering the relationship between the number of the initial conditions and the differential order of the design equations, factoring in scale invariance of the design equations.

This study investigates the effect of frame intervals on the accuracy of the Lucas–Kanade optical flow estimates for high-frame-rate (HFR) videos, with a view to realizing accurate HFR-video-based optical flow estimation. For 512 512 pixels videos of patterned objects moving at different speeds and captured at 1000 frames per second, the averages and standard deviations of the estimated optical flows were determined as accuracy measures for frame intervals of 1–40 ms. The results showed that the accuracy was highest when the displacement between frames was around 0.6 pixel/frame. This common property indicates that accurate optical flow estimation for HFR videos can be realized by varying frame intervals according to the motion field: a small frame interval for high-speed objects and a large frame interval for low-speed objects.

In a wireless sensor network based on the gradient sinking model, unbalanced energy consumption is an inherent problem and can significantly reduce the network lifetime. In this letter, we propose a subcorona-based scheme to analyze the amount of received data and energy consumption of nodes on gradient sinking model. We then design an algorithm to compute the energy consumption of nodes in different subcoronas. Simulation results indicate the correctness of our proposed algorithm.

The purpose of this study is to propose an advanced phase-based optical flow method with improved tracking accuracy for motion flow. The proposed method is mainly based on adaptive bilateral filtering (ABF) and Gabor based spatial filtering. ABF aims to preserve the maximum boundary information of the original image, while the spatial filtering aims to accurately compute the local variations. Our method tracks the optical flow in three stages. Firstly, the input images are filtered by using ABF and a spatial filter to remove noises and to preserve the maximum contour information. The component velocities are then computed based on the phase gradient of each pixel. Secondly, irregular pixels are eliminated, if the phase differences are not linear over the image frames. Lastly, the entire velocity is derived by integrating the component velocities of each pixel. In order to evaluate the tracking accuracy of the proposed method, we have examined its performance for synthetic and realistic images for which the ground truth data were known. As a result, it was observed that the proposed technique offers higher accuracy than the existing optical flow methods.

In order to implement multidimensional scaling (MDS) efficiently, we propose a new method named “global mapping analysis” (GMA), which applies stochastic approximation to minimizing MDS criteria. GMA can solve MDS more efficiently in both the linear case (classical MDS) and non-linear one (e.g., ALSCAL) if only the MDS criteria are polynomial. GMA separates the polynomial criteria into the local factors and the global ones. Because the global factors need to be calculated only once in each iteration, GMA is of linear order in the number of objects. Numerical experiments on artificial data verify the efficiency of GMA. It is also shown that GMA can find out various interesting structures from massive document collections.

Space shift keying (SSK), conveying data symbols via only the antenna indices, is a new modulation technique for low-complexity implementation of multiple-input multiple-output (MIMO) systems. SSK can be combined with the orthogonal frequency division multiplexing (OFDM) technique to improve the capacity and reliability of transmission. However, the SSK MIMO-OFDM systems also inherit from OFDM systems the drawback of a high peak-to-average power ratio (PAPR) of the transmitted signal. To overcome this problem, in this paper, the special information-conveying mode of SSK is utilized and a fast-converged gradient-based PAPR reduction method which exploits the phase freedom of the transmitted SSK MIMO-OFDM signal in the frequency domain is proposed. Simulation results show that the proposed improved gradient-based method (GBM) achieves a superior PAPR performance, as compared to the promising discrete phase set based schemes such as selected mapping method (SLM) and flipping method. Besides the considerable PAPR reduction, the improved GBM also enjoys other advantages such as low complexity, and neglectable performance loss.

This paper presents a novel method for reconstructing 3D geometry of camera motion and human-face model from a video sequence. The approach combines the concepts of Powell's line minimization with gradient descent. We adapted the line minimization with bracketing used in Powell's minimization to our method. However, instead of bracketing and searching deep down a direction for the minimum point along that direction as done in their line minimization, we achieve minimization by bracketing and searching for the direction in the bracket which provides a lower energy than the previous iteration. Thus, we do not need a large memory as required by Powell's algorithm. The approach to moving in a better direction is similar to classical gradient descent except that the derivative calculation and a good starting point are not needed. The system's constraints are also used to control the bracketing direction. The reconstructed solution is further improved using the Levenberg Marquardt algorithm. No average face model or known-coordinate markers are needed. Feature points defining the human face are tracked using the active appearance model. Occluded points, even in the case of self occlusion, do not pose a problem. The reconstructed space is normalized where the origin can be arbitrarily placed. To use the obtained reconstruction, one can rescale the computed volume to a known scale and transform the coordinate system to any other desired coordinates. This is relatively easy since the 3D geometry of the facial points and the camera parameters of all frames are explicitly computed. Robustness to noise and lens distortion, and 3D accuracy are also demonstrated. All experiments were conducted with an off-the-shelf digital camera carried by a person walking without using any dolly to demonstrate the robustness and practicality of the method. Our method does not require a large memory or the use of any particular, expensive equipment.

Complex-valued Associative Memory (CAM) is an advanced model of Hopfield Associative Memory. The CAM is based on multi-state neurons and has the high ability of representation. Lee proposed gradient descent learning for the CAM to improve the storage capacity. It is based on only the phases of input signals. In this paper, we propose another type of gradient descent learning based on both the phases and the amplitude. The proposed learning method improves the noise robustness and accelerates the learning speed.

We present a review of several types of microwave antennas made of metamaterials, including the resonant electrically small antennas, metamaterial-substrate patch antennas, metamaterial flat-lens antennas, and Luneburg lens antennas. In particular, we propose a new type of conformal antennas using anisotropic zero-index metamaterials, which have high gains and low sidelobes. Numerical simulations and experimental results show that metamaterials have unique properties to design new antennas with high performance.

This paper presents a robust framework of human-computer interaction from the hand gesture vision in the presence of realistic and challenging scenarios. To this end, several novel components are proposed. A hybrid approach is first proposed to automatically infer the beginning position of hand gestures of interest via jointly optimizing the regions given by an offline skin model trained from Gaussian mixture models and a specific hand gesture classifier trained from the Adaboost technique. To consistently track the hand in the context of using kernel based tracking, a semi-supervised feature selection strategy is further presented to choose the feature subspaces which appropriately represent the properties of offline hand skin cues and online foreground-background-classification cues. Taking the histogram of oriented gradients as the descriptor to represent hand gestures, a soft-decision approach is finally proposed for recognizing static hand gestures at the locations where severe ambiguity occurs and hidden Markov model based dynamic gestures are employed for interaction. Experiments on various real video sequences show the superior performance of the proposed components. In addition, the whole framework is applicable to real-time applications on general computing platforms.

The conjugate gradient-fast multipole method (CG-FMM) is one of the powerful methods for analysis of large-scale electromagnetic problems. It is also known that CPU time and computer memory can be reduced by CG-FMM but such computational cost of CG-FMM depends on shape and electrical properties of an analysis model. In this paper, relation between the number of multipoles and number of segments in each group is derived from dimension of segment arrangement in four typical wiregrid models. Based on the relation and numerical results for these typical models, the CPU time per iteration and computer memory are quantitatively discussed. In addition, the number of iteration steps, which is related to condition number of impedance matrix and analysis model, is also considered from a physical point of view.

Electrostatic discharge (ESD) events due to metal objects electrified with low voltages give a fatal electromagnetic interference to high-tech information equipment. In order to elucidate the mechanism, with a 6-GHz digital oscilloscope, we previously measured the discharge current due to collision of a hand-held metal piece from a charged human body, and gave a current calculation model. In this study, based on the calculation model, a method was presented for deriving a gap potential gradient from the measured discharge current. Measurements of the discharge currents were made for charge voltages from 200 V to 1000 V. The corresponding potential gradients were estimated, which were validated in comparison with an empirical formula based on the Paschen's law together with other researcher's experimental results.

We propose the multi-domain adaptive learning that enables us to find a point meeting possibly time-varying specifications simultaneously in multiple domains, e.g. space, time, frequency, etc. The novel concept is based on the idea of feasibility splitting -- dealing with feasibility in each individual domain. We show that the adaptive projected subgradient method (Yamada, 2003) realizes the multi-domain adaptive learning by employing (i) a projected gradient operator with respect to a ‘fixed’ proximity function reflecting the time-invariant specifications and (ii) a subgradient projection with respect to ‘time-varying’ objective functions reflecting the time-varying specifications. The resulting algorithm is suitable for real-time implementation, because it requires no more than metric projections onto closed convex sets each of which accommodates the specification in each domain. A convergence analysis and numerical examples are presented.

In this paper, we propose a novel stochastic gradient algorithm for efficient adaptive filtering. The basic idea is to sparsify the initial error vector and maximize the benefits from the sparsification under computational constraints. To this end, we formulate the task of algorithm-design as a constrained optimization problem and derive its (non-trivial) closed-form solution. The computational constraints are formed by focusing on the fact that the energy of the sparsified error vector concentrates at the first few components. The numerical examples demonstrate that the proposed algorithm achieves the convergence as fast as the computationally expensive method based on the optimization without the computational constraints.

In and we have presented a simulation-based algorithm for optimizing the average reward in a parameterized continuous-time, finite-state semi-Markov Decision Process (SMDP). We approximated the gradient of the average reward. Then, a simulation-based algorithm was proposed to estimate the approximate gradient of the average reward (called GSMDP), using only a single sample path of the underlying Markov chain. GSMDP was proved to converge with probability 1. In this paper, we give bounds on the approximation and estimation errors for GSMDP algorithm. The approximation error of that approximation is the size of the difference between the true gradient and the approximate gradient. The estimation error, the size of the difference between the output of the algorithm and its asymptotic output, arises because the algorithm sees only a finite data sequence.