Advanced Materials Research Vols. 846-847

Abstract: Hidden Markov model HMM (1) is one of the important approaches for information extraction. In this paper, a model of the improved first-order hidden Markov HMM (2) is proposed. In the HMM (2), the output probability of the observation is not only dependent on the current state of the model, but also dependent on the previous state of the current state of the model. The algorithm of the ML and the algorithm of the Viterbi are analyzed. At last, experiments show that the HMM (2) is more precise than the HMM (1).

Abstract: In flow visualization, the streamline has always been used to display the vector fields, but only using streamlines cannot reveal the nature of flowing intuitively. The dynamic glyphs technology is an important method to show flowing 3D vector fields. However, existing methods of dynamic glyphs are hard to convey the global patterns and local entities at the same time. In this paper, we present a practical method for dynamic visualization of 3D vector fields which combine dynamic glyphs and streamlines, which are rendered respectively according to the information of 3D vector fields. Our method integrates virtues of these two usual techniques, and provides high performance visualization of flowing 3D vector fields directly. Experiments with practical datasets certify that our method could not only depict the local and the global patterns, but more importantly reveal the dynamic nature of flowing 3D vector data.

Abstract: The field of algorithmically assessing the 3D quality of experience is an extremely challenging one. In this paper a no-reference algorithm is proposed to assess the comfort associated with viewing stereo videos. The proposed measure of 3D quality of experience is shown to correlate well with human perception of quality, and extracts statistical features from disparity and disparity gradient maps. And the measure utilizes these spatial features along with motion compensated disparity differences to predict quality of experience.

Abstract: Mining partial orders from sequence data is an important data mining task with broad applications. As partial orders mining is a NP-hard problem, many efficient pruning algorithm have been proposed. In this paper, we improve a classical algorithm of discovering frequent closed partial orders from string. For general sequences, we consider items appearing together having equal chance to calculate the detecting matrix used for pruning. Experimental evaluations from a real data set show that our algorithm can effectively mine FCPO from sequences.

Abstract: Natural language processing is an important subject in the field of artificial intelligence; one of the key issues is the quantification research on the fuzzy semantics, and the fuzziness of natural language is mostly reflected on the adverbs. This paper presented one new method to quantify the English adverbs. After briefly introducing the Medium Logic (ML) and Measure of Medium Truth Degree (MMTD), commendatory and derogatory were qualitatively described with logical predicates C and C respectively; the truth degree related to C or C of an adverb could be quantitatively calculated by the model proposed based on the existed research achievement of linguistics in this paper. The given computing of frequency adverb and degree adverb shows that above methods to process fuzzy semantics of adverbs is effective.

Abstract: The common point coordinate precision affect accuracy of coordinate transformation. In order to overcome the individual common point coordinate precision is too low or the error is large under some condition, a new method of high-precision coordinate transformation based on bursa model was discussed, which used the theory of robust resistance estimation to solve and calculate the high-precision coordinate transformation parameters, finally achieved the purpose of high-precision coordinate transformation. By calculation and comparison with experimental date, the result show that the method is practical and effective, so it is completely feasible in application surveying engineering.

Abstract: Presently, most of the PSO algorithms for eigenvalues computation take the characteristic polynomial as its fitness function, while for large matrix; it needs a lot of computation for its determination. In order to improve the efficiency of arithmetic operations, a new definition of fitness function, the fitness function calculation of which is much less than the calculation of the determination, can greatly improve the efficiency of algorithms. Finally, several different types of matrix operator cases further validate the effectiveness of the obtained results.

Abstract: In the paper, we have a study on reconstruction technology from single image, particularly on geometric constraint based methods. We analyze several important geometry constraint relationships in the image and derive corresponding function expression. In this way, the depth of each image point can be gained relatively accurately. With the depth of object points, we fulfilled model reconstruction from a single image. We call the single image reconstruction method as sequential method which need calculate depth of object points one by one. In the experiment, the results prove the visual effect of the algorithm is satisfactory.

Abstract: The sentence transformation from phonetic to word is a very critical part of the input method. When the input method can not find the candidate through a dictionary word directly, it needs to obtain the results desired by the user through sentence transformation. In this paper, based on recursive enumeration k-best decoding algorithm is used in the input method sentence transformation , with the language model, to get the k-optimal transformation results . Experimental results show that in the input method application environment, based on a recursive enumeration k-best decoding algorithm decoding efficiency significantly better than the deletion algorithm which is the baseline.

Abstract: The degree-constrained Euclidean Steiner minimal tree problem was discussed based on the Euclidean Steiner minimal tree with each original point being added with a degree constraint. The property of the problem was analyzed and the implementation process of solving the problem by using the ant algorithm was presented.The algorithms is coded in Delphi and run on the Windows XP environment.