The principle of target tracking and data fusion techniques are discussed. To resolve high uncertainty that exists in sensors of mobile robots, one multi-sensor data fusion algorithm is presented. The algorithm is bas...
The principle of target tracking and data fusion techniques are discussed. To resolve high uncertainty that exists in sensors of mobile robots, one multi-sensor data fusion algorithm is presented. The algorithm is based on particle filter techniques, fuses the information coming from multiple sensors and merges different state space models. So it can be used to eliminate system and measurement noise and estimate value of position and headings of mobile robot. On simulation experiments, we compare different cases such as single sensors and multi-sensor data fusion, the results demonstrate the feasibility and effectiveness of this algorithm and exhibits good tracking performance.
In this paper, based on Baldwin effect, an improved clonal selection algorithm, Baldwin clonal selection algorithm, termed as BCSA, is proposed to deal with complex multimodal optimization problems. BCSA evolves and i...
详细信息
In this paper, based on Baldwin effect, an improved clonal selection algorithm, Baldwin clonal selection algorithm, termed as BCSA, is proposed to deal with complex multimodal optimization problems. BCSA evolves and improves antibody population by three operations: clonal proliferation operation, Baldwinian learning operation and clonal selection operation. By introducing Baldwin effect, BCSA can make the most of experience of antibodies, accelerate the convergence, and obtain the global optimization quickly. In experiments, BCSA is tested on four types of functions and compared with the clonal selection algorithm and other optimization methods. Experimental results indicate that BCSA achieves a good performance, and is also an effective and robust technique for optimization.
Support vector machine, a universal method for learning from data, gains its development based on statistical learning theory. It shows many advantages in solving nonlinearly small sample and high dimensional problems...
详细信息
Support vector machine, a universal method for learning from data, gains its development based on statistical learning theory. It shows many advantages in solving nonlinearly small sample and high dimensional problems of pattern recognition. Only a part of samples or support vectors (SVs) plays an important role in the final decision function. But SVs could not be obtained in advance until a quadratic programming is performed. In this paper, we use K-nearest neighbour method to extract a boundary vector set which may contain SVs. The number of the boundary set is smaller than the whole training set. Consequently it reduces the training samples, speeds up the training of support vector machine.
A method for multi-classifier ensemble of Support Vector Machine ensemble (SVMs) and Kernel Matching Pursuit Ensemble (KMPs) is proposed. Support Vector Machine has advantage in solving classification problem of high ...
详细信息
A method for multi-classifier ensemble of Support Vector Machine ensemble (SVMs) and Kernel Matching Pursuit Ensemble (KMPs) is proposed. Support Vector Machine has advantage in solving classification problem of high dimension and small size dataset, and Kernel Matching Pursuit has almost classified performance and the more sparsely solution as comprised with the SVM. So the SVM and the KMP are mix boosted in this paper, which can decrease generalization errors of the single classifier ensemble and improve ensemble classification accuracy by increasing diversity between ensemble individuals. The experiments show that the proposed method can shorten running time and improve classification accuracy compared with individual SVMs or KMPs.
A progressive image compression scheme is investigated using reversible integer discrete cosine transform (RDCT) which is derived from the matrix factorization theory. Previous techniques based on DCT suffer from bad ...
详细信息
A progressive image compression scheme is investigated using reversible integer discrete cosine transform (RDCT) which is derived from the matrix factorization theory. Previous techniques based on DCT suffer from bad performance in lossy image compression compared with wavelet image codec. And lossless compression methods such as IntDCT, I2I-DCT and so on could not compare with JPEG-LS or integer discrete wavelet transform (DWT) based codec. In this paper, lossy to lossless image compression can be implemented by our proposed scheme which consists of RDCT, coefficients reorganization, bit plane encoding, and reversible integer pre- and post-filters. Simulation results show that our method is competitive against JPEG-LS and JPEG2000 in lossless compression. Moreover, our method outperforms JPEG2000 (reversible 5/3 filter) for lossy compression, and the performance is even comparable with JPEG2000 which adopted irreversible 9/7 floating-point filter (9/7F filter).
Based on the theory of clonal selection in immunology, by introducing Baldwin effect, an improved clonal selection algorithm, termed as Baldwin clonal selection algorithm (BCSA), is proposed to solve the optimal appro...
详细信息
Based on the theory of clonal selection in immunology, by introducing Baldwin effect, an improved clonal selection algorithm, termed as Baldwin clonal selection algorithm (BCSA), is proposed to solve the optimal approximation of linear systems. For engineering computing, the novel algorithm adopts three operations to evolve and improve the population: clonal proliferation operation, Baldwinian learning operation and clonal selection operation. The experimental study on the optimal approximation of a stable linear system and an unstable one show that the approximate models searched by the new algorithm have better performance indices than those obtained by some existing algorithms including the differential evolution algorithm, multi-agent genetic algorithm and artificial immune response algorithm.
In this paper, we introduce Lamarckian learning theory into the clonal selection algorithm and propose a sort of Lamarckian clonal selection algorithm, termed as LCSA. The major aim is to utilize effectively the infor...
详细信息
In this paper, we introduce Lamarckian learning theory into the clonal selection algorithm and propose a sort of Lamarckian clonal selection algorithm, termed as LCSA. The major aim is to utilize effectively the information of each individual to reinforce the exploitation with the help of Lamarckian local search. Recombination operator and tournament selection operator are incorporated into LCSA to further enhance the ability of global exploration. We compared LCSA with the clonal selection algorithm (CSA) in solving twenty benchmark problems to test the performance of LCSA. The results demonstrate that LCSA is effective and efficient in solving numerical optimization problems.
An efficient image denoising algorithm is introduced. Firstly, image pixels are classified into noisy pixels and noise-free pixels by four directional operators. Then an adaptive weighted median filter is designed to ...
详细信息
An efficient image denoising algorithm is introduced. Firstly, image pixels are classified into noisy pixels and noise-free pixels by four directional operators. Then an adaptive weighted median filter is designed to remove and restore the detected noisy pixels and keep the noise-free ones unchanged. Experimental results indicate that the proposed algorithm preserves image details well while removing impulsive noise efficiently, and its filtering performance is significantly superior to the classical median filter and some other typical and recently developed improved median filters.
Aero-optic effects cause distortions, including blurring, vibration, deformation and spatial shifting, of the objects in the image obtained by the infra-red sensor. Contributions of this paper are in the following two...
详细信息
ISBN:
(纸本)9781424439027
Aero-optic effects cause distortions, including blurring, vibration, deformation and spatial shifting, of the objects in the image obtained by the infra-red sensor. Contributions of this paper are in the following two aspects. First, the correctness of the theoretical point spread function (PSF) representing the aero-optic effects, which had been derived in our previous research, is validated experimentally. Second, in order to restore the aero-optically degraded images, an improved Landweber iteration method is proposed, where, instead of being fixed, the relaxation factor is updated adaptively at each iteration. Experiments have been carried out and results demonstrate that the proposed method introduces improved restoration results with better convergence.
This article consists of a collection of slides from the author's conference presentation. Some of the specific conclusions presented/discussed include: Rebalanced architecture to workload trends; Scaled from 128 ...
详细信息
This article consists of a collection of slides from the author's conference presentation. Some of the specific conclusions presented/discussed include: Rebalanced architecture to workload trends; Scaled from 128 to 240 processors; Hardware manages thousands of threads; Zero software overhead; Hides huge latencies; High achieved utilization; Natively Scalar; No swizzling or vectorization overhead; Coalescing for high bandwidth memory I/O; Software architecture allows 2X scaling on customer C code with no modification.
暂无评论