Advanced Materials Research Vols. 846-847

Abstract: Online-handwriting of Uyghur word categorization (UWC) based on improved K-NN is studied in this paper. Firstly, features are selected by correlation analysis method (CAM), and correlation is measured by linear correlation coefficient (LCC). Secondly, K-NN algorithm is improved and classifier is established. Finally, the classifier is applied in UWC. The experiment results show that the method has good performance on online handwriting of UWC.

Abstract: In this paper, linking with the basic principle of FCM (Fuzzy c-means clustering) algorithm, on the basis of theory research, the segmented partitions emerge when the state of the lives reaches an equilibrium. The artificial life approach is promising in image processing because it is inherently parallel and coincides with the self-governing biological process. Then combined with intensity-texture-position feature space in order to produce connected regions shown in the image, the final segmentation result is achieved at last. The experiment results prove that in the view of the sport image segmentation, this algorithm provides fast segmentation with high perceptual segmentation quality.

Abstract: This paper presents a novel correlation-based tracking algorithm, which is a generalization of the dynamic template correlation regional tracking (DTCRT) algorithmthe template is dynamic and the correlation is regional, and in particular the target image quality is not always good. The algorithm does not make use of the point-to-point multiplication and then summarizing as the traditional algorithm does. Instead, the DTCRT algorithm introduced. This fact helps to overcome several shortcomings of the point-to-point multiplication and then summarizing. The target is represented by the correlation coefficients. If the correlation coefficient is less than a threshold one can judge then the spike is defected. Since the matching function is a very complex function, the DTCRT algorithm is used to optimize it. Experiments demonstrate the effectiveness, significance and computation efficiency of the proposed railway spike tracking method in real-time identification.

Abstract: In recent years, the target of local feature extraction and recognition technology is a hot research topic in digital image processing and computer vision field.In target recognition, pest undetected phenomenon often happens due to occlusion, once the conditions are favorable, pests quickly multiply and spread, causing great harm to the stored product.we established a common stored product pests image database, proposed feature extraction and matching algorithms, based on the combination of improved SURF and multi-resolution histogram.Pests specified in the database for target identification, especially when the pest is partially obscured, or the shape and size changes,by finding matching number of feature points the description of the target for accurate identification .This method is efficient, time-saving,provides an important scientific basis for the entry-exit inspection and quarantine in rapid identification of stored product pests.It is more conducive to the subsequent classification.

Abstract: This design is based on WSSUS channel model. We achieved a real-time signal processing channel emulator which used the channel parameters provided by the DRM criteria. The article first introduces the WSSUS channel model, and then bases DRM criteria to determine the model parameters. Finally, through a structured design, we realized the model HF channel emulator by C language programming.

Abstract: This paper does a performance comparison and evaluation to the CF algorithm based on the cosine similarity, the correlation similarity and project rating, and analyzes and researches its application, facing problems, solutions in the personalization recommendation system.

Abstract: A comprehensive analysis on a large amount of ship equipment consumption data accumulated over the years is achieved through the establishment of data warehouse, online analytical processing, regression analysis, cluster analysis, etc. by means of data mining. The analysis results present important references for equipment guarantee department in terms of equipment preparation and carrying, etc. and provide the comprehensive analysis and utilization on massive ship maintenance support data with technical means.

Abstract: In the field of bioinformatics visualization, integrating software and data in different levels is the development trend. This paper presents an integration framework for biomolecular structure and genome sequences visualization. The framework can effectively support the data and software interoperability of biomolecular structure / genome sequences visualization. Based on the framework, we developed an integrated visualization system, which provides some new comprehensive visualization functions. Preliminary trial showed that the framework has a good prospect in the research of bioinformatics.

Abstract: in recent years, the network has placed everywhere of people living and production under the network technology development. It is one of the necessities in our daily life. The network information technology expands peoples view, closing the distance, change the life and working mode, and develops printing industry innovation. After understanding about the requirements about printing picture characteristic and picture detection, the researchers develop new printing picture detection. This is the full-screen detection. Although this technology obtains well application, it's still in a fielding period and need further perfection. This article pays attention to implementation plan and control strategy of full-screen detection.

Abstract: This paper carries out research on life detection by micro Doppler.Micro motion parameters can be estimated through extracting the micro Doppler signatures.The article establishes the human body model and radar echo model. Then STFT,CWT and generalized S transform are analyzed and compared to extract micro Doppler signatures,and improvement of generalized S transform is carried out to enhance its frequency aggregation and noise suppression ability.Then principal component analysis and support vector machine are studied.By extracting principal components from the micro Doppler spectrum as input of support vector machine,classification and identification is complished. The simulation results show the improved generalized S transform has better recognition accuracy in noise condition.