In this paper, we propose an approach to fuse the color and infrared images for visual tracking. The contribution of this paper is twofold: First, we use the covariance feature to construct the likelihood function und...
详细信息
In this paper, we propose an approach to fuse the color and infrared images for visual tracking. The contribution of this paper is twofold: First, we use the covariance feature to construct the likelihood function under the framework of particle filter. This likelihood captures the spatial and statistical properties as well as their correlation within representation of covariance. Secondly, different from the existing fusion approaches, our approach automatically realizes the fusion by sequential belief propagation, which uses message passing scheme to exchange information between color and infrared image. The performance of the proposed approach is evaluated using real visual tracking examples.
In order to generate natural posture and motion of virtual human-like figures, damped least squares inverse kinematics method is modified. Physical rules of human being like joint limits, joint weights and comfortable...
详细信息
In order to generate natural posture and motion of virtual human-like figures, damped least squares inverse kinematics method is modified. Physical rules of human being like joint limits, joint weights and comfortable criteria are introduced to the design of damping factors for the improved damped least squares solution. The proposed method performs well on guaranteeing joint limit avoidance and producing natural-looking postures. This new scheme is successfully implemented and tested for real-time control of a seven-degree-of-freedom virtual human skeletal upper limb. Experiment results show that the improved solution is more robust and stable than the original damped least squares method.
We proposed a fully-software distributed failure diagnosis system for vehicles based on the TH-OSEK real-time embedded OS platform we previously developed. The diagnosis system puts all the ECUs into a virtual logical...
详细信息
ISBN:
(纸本)9781424418480
We proposed a fully-software distributed failure diagnosis system for vehicles based on the TH-OSEK real-time embedded OS platform we previously developed. The diagnosis system puts all the ECUs into a virtual logical ring and uses the MR(Maintain Ring) algorithm and OL(Off Line) algorithm to detect a faulted ECU and isolate it without destroying the structure of the logical ring. When a faulted ECU is recovered, with the proposed algorithms the system can also add it to the logical ring by updating the predecessor and successor of every node in the ring in time. The experiment result on the TH-OpenECU platform is also presented which shows that the system works well and usefulness for diagnosing the faults of vehicles.
Conventional cost functions of adaptive filtering are usually related to the errorpsilas dispersion, such as errorpsilas moments or errorpsilas entropy, but neglect the shape aspects (peaks, kurtosis, tails, etc.) of ...
详细信息
Conventional cost functions of adaptive filtering are usually related to the errorpsilas dispersion, such as errorpsilas moments or errorpsilas entropy, but neglect the shape aspects (peaks, kurtosis, tails, etc.) of the error distribution. In this work, we propose a new notion of filtering (or estimation) in which the errorpsilas probability density function (PDF) is shaped into a desired one. As PDFs contain all the probabilistic information, the proposed method can be used to achieve the desired error variance or error entropy, and is expected to be useful in the complex signal processing and learning systems. In our approach, the information divergence between the actual errors and the desired errors is used as the cost function. By kernel density estimation, we derive the associated stochastic gradient algorithm for the finite impulse response (FIR) filter. Simulation results emphasize the effectiveness of this new algorithm in adaptive system training.
A machine learning approach to predict turning points for chaotic time series was proposed through incorporating chaotic analysis into ensemble artificial neural network (ANN) modeling. The EM-like parameter learning ...
详细信息
A machine learning approach to predict turning points for chaotic time series was proposed through incorporating chaotic analysis into ensemble artificial neural network (ANN) modeling. The EM-like parameter learning algorithm for ensemble ANN model was presented. We then gave a new GA-based threshold optimization procedure using out-of-sample validation. The proposed approach was demonstrated on the benchmark chaotic time series like Mackey-Glass system. Our experimental results show significant improvement in performance over ANN model alone.
We study the unsorted database search problem with items N from the viewpoint of unitary discrimination. Instead of considering the famous O(N) Grover bounded-error algorithm for the original problem, we seek the resu...
详细信息
We study the unsorted database search problem with items N from the viewpoint of unitary discrimination. Instead of considering the famous O(N) Grover bounded-error algorithm for the original problem, we seek the results for the exact algorithms, i.e., those that succeed with certainty. Under the standard oracle model ∑j(−1)δτj|j⟩⟨j|, we demonstrate a tight lower bound 23N+o(N) of the number of queries for any parallel scheme with unentangled input states. With the assistance of entanglement, we obtain a general lower bound 12(N−N). We provide concrete examples to illustrate our results. In particular, we show that the case of N=6 can be solved exactly with only two queries by using a bipartite entangled input state. Our results indicate that in the standard oracle model the complexity of the exact quantum search with one unique solution can be strictly less than that of the calculation of the OR function.
We present a complete characterization for the local distinguishability of orthogonal 2⊗3 pure states except for some special cases of three states. Interestingly, we find there is a large class of four or three state...
详细信息
We present a complete characterization for the local distinguishability of orthogonal 2⊗3 pure states except for some special cases of three states. Interestingly, we find there is a large class of four or three states that are indistinguishable by local projective measurements and classical communication (LPCC), but can be perfectly distinguished by LOCC. That indicates the ability of LOCC for discriminating 2⊗3 states is strictly more powerful than that of LPCC, which is strikingly different from the case of multiqubit states. We also show that classical communication plays a crucial role for local distinguishability by constructing a class of m⊗n states which require at least 2min{m,n}−2 rounds of classical communication in order to achieve a perfect local discrimination.
Semi-supervised image segmentation is an important issue in many image processing applications, and has been a popular research area recently, the most popular are graph-based methods. However, parameter selection in ...
详细信息
Semi-supervised image segmentation is an important issue in many image processing applications, and has been a popular research area recently, the most popular are graph-based methods. However, parameter selection in these methods is still largely heuristic. In this paper, we introduce distance metric learning into graph-based semi-supervised segmentation to automatically obtain good results for images with different appearances. We first derive the optimization problem with respect to the distance metric as well as the segmentation labels, and use gradient descent method to find a local optimum solution. Experiments on general images and the fungal disease analysis application have shown that our method provides a steady performance under casual user annotations and different image appearances.
Automatic evaluation of perceptual similarity is crucial for music retrieval. However, previous works mainly focused on the similarity of timbre and rhythm but not the musical pattern of a song, such as melody and cho...
详细信息
Automatic evaluation of perceptual similarity is crucial for music retrieval. However, previous works mainly focused on the similarity of timbre and rhythm but not the musical pattern of a song, such as melody and chord. In this paper, we propose a new feature, chroma histogram, to summarize the musical pattern and use a transposition-invariant matching method to compare two chroma histograms. Experiment results demonstrate the efficiency of this method in measuring the similarity of musical pattern.
In this paper we discuss the issue of classifiers combined with Histogram of Oriented Gradients (HOG) descriptors for human detection. And we present a method that combines AdaBoost learning with HOG descriptors. The ...
详细信息
In this paper we discuss the issue of classifiers combined with Histogram of Oriented Gradients (HOG) descriptors for human detection. And we present a method that combines AdaBoost learning with HOG descriptors. The weak learners used in our algorithm are based on weighted Modified Quadratic Discriminant Functions (MQDF) which is a parametric model. We evaluate our algorithm on the INRIA person dataset. And the experimental results show that our approach achieves a comparable performance with the state of art methods both on accuracy and speed.
暂无评论