Advanced Materials Research Vols. 1049-1050

Abstract: Data mining aims to excavate new knowledge from existing information. When it comes to test mining, a better way is to take the context into account In this study we present text mining procedures based on a neural network framework in order to identify indicative factors in the form of keywords within the medical record narratives. These keywords and their weight/value suggest an innovative way for justifying a CT scan request. Our purpose is to extend the reach of diagnosis beyond traditional processing of clinical data towards an efficient utilization of the narratives in medical records.

Abstract: According to CT and MR image features, we presented the medical image fusion algorithm based on adaptive Gaussian wavelet, which uses gray level co-occurrence matrix to modify space coefficients adaptively. Compared with other wavelet algorithm by simulation, the results proved that the proposed algorithm obtained valid information in a highly automated manner, and without human intervention or prior information.

Abstract: Image denoising method based on K-SVD self-learning dictionary can effectively filter Gaussian white noise in an image and retain image details and texture information. This paper proposed an improved denoising algorithm which was based on K-SVD algorithm, but with adaptive dictionary size. Depending on the complexity of image content and the noise level, our algorithm determines the size of the dictionary adaptively. Experimental results show that proposed algorithm can reduce the number of entries of the dictionary significantly for the simple images, and increase the number of entries for the complex images. Both the efficiency and the denoising performance are improved compared with original K-SVD algorithm.

Abstract: In order to simulate the plant branch structure in three-dimensional space, and extract the growth more efficiently, this paper presented a new method to simulate the structure of plant branch based on quasi binary-tree structure and three-dimensional L system. The results of the actual trees simulation shows that this method can describe the plants branch structure efficiently and provide a new way for the simulation of plants.

Abstract: To optimize the parameters of LS-SVM effectively, an improved Particle Swarm Optimization (PSO) algorithm is proposed to select the optimal parameters combination. For the improvement of the precocity in PSO algorithm, an multi-particles sharing strategy is introduced in simple PSO algorithm to enhance the convergence. The simulation indicates that the proposed PSO algorithm has a better selection on LS-SVM parameters.

Abstract: In order to improve the performance of nonlinear modeling, a Hopfield neural network modeling method based on Subset Kernel Principal Components Analysis (SubKPCA) with Fuzzy C-Means Clustering (FCMC) is proposed. The simulation result shows that, the performance of the proposed method is better than that of Hopfield neural network based on KPCA. It also is effective and feasible to establish the model for the estimation of missing flight data.

Abstract: Time-table Problem of universities is a many factor of the global optimization problem. In this paper, according to the characteristics of time-table problem, an improve Cuckoo Search Algorithm was used to solve the Time-table Problem, adopting the code rule of randomized key representation based on the smallest position value, and then the design scheme of time-table problem of universities based on improved cuckoo search algorithm was expounded through studying influence factors of time-table problem of universities. Finally,the result shows the algorithm is feasible and effective.

Abstract: To improve the modeling performance of Recurrent Wavelet Neural Network (RWNN), a training algorithm based on Immune Evolving Algorithm (IEA) is proposed. In the process of RWNN training, IEA is mainly used to optimize the connection weight, translating and scaling parameter. The experiment result on Duffing chaotic time series shows that the proposed RWNN training algorithm has a good prediction capability in the field of nonlinear modeling.

Abstract: In this paper, according to the characteristics of nonlinear equations. A Novel Firefly Algorithm was used to solve the nonlinear equations problem, the algorithm was experimented and the experimental results show that the new algorithm to be successful in locating multiple solutions and better accuracy. The experimental result demonstrates that the Improve firefly algorithm can get better solutions to nonlinear equations.

Abstract: Object extraction, which aims to accurately separate a foreground object from its background in still images, plays an important role in many computer vision applications. An interactive object extraction method based on MSRM (maximal similarity based region merging) is presented in this paper. We can manually mark the target and background only one time in any one image of the image sequence to obtain the object extraction result of the image sequence. Compared to currently used method based on graph cut algorithm that manually marks the target and background on all the images one by one to get the object extraction result, our method is more efficient and the result is as precious as with other methods.