Time filter

Source Type

Wang J.,Soochow University of China | Sun Y.,Soochow University of China | Sun Y.,Jiangsu Provincial Key Laboratory of Computer Information Processing Technology | Wang C.,Soochow University of China
Journal of Soochow University Engineering Science Edition

Similarity computing in patent information analysis is used widely. The traditional way of text similarity computing based on vector space model have two defects. One is that limitation put on word semantic by context was not considered enough, making it hard to effectively distinguish word semantic in diverse contexts. The other is that its use of dictionaries as foundation of similarity computing carried domain correlation, making it incapable of offering effective semantic comprehension in specific domain. Domain ontology is employed in this paper as the basis for text similarity computing, and sense disambiguation is applied to resolve the influence put on word semantic by context. Besides, maximum match of bipartite graph is used to carry out similarity computing. The experience result showed that the accuracy of this method enjoys improvements compared with the traditional way. Source

Zhao H.,Hunan University | Yang Y.,Hunan University | Lu Z.,Hunan University | Yu F.,Jiangsu Provincial Key Laboratory of Computer Information Processing Technology
Journal of Computers (Finland)

The traditional Bayesian estimator of short-time spectral amplitude is based on the minimization of the squared-error cost function under the common Gaussian probability density function (pdf). The Gaussian distribution, however, is not the optimal probability distribution. To overcome this phenomenon, we considered to replace the traditional distribution hypothesis of spectral amplitude of speech in this paper. More precisely, we proposed a β -order perceptive Bayesian spectral amplitude estimator which incorporated the assumption of Super-Gaussian chi-distributed spectral amplitude. The new weighting function incorporated the perceptive property as well as the different importance of the spectral valley and peak. Experiments showed that the proposed estimator can achieve a more significant noise reduction and yield a better spectral estimation over the most of latest enhancement algorithms. © 2012 ACADEMY PUBLISHER. Source

Jiang X.,Suzhou University | Sun Y.,Suzhou University | Sun Y.,Jiangsu Provincial Key Laboratory of Computer Information Processing Technology
Journal of Soochow University Engineering Science Edition

It is one of the most popular and challenging subjects in recently years to improve visualization and efficiency of moving fluid in scientific researches, engineering projects, computer games, film-video works, and visual simulation. This thesis proposes a kind of betterment algorithm that applies the spline interpolation called Catmull-Rom in cell subdivision of squares, so as to conduct visual simulation in 2D-wave surface images. Such problems as roughness, too many bend points and complicated computation caused by cell subdivision of moving fluid surface can be solved by the algorithms to get smooth images on the premise of ensuring computation efficiency. The problem of ambiguity can get resolution by virtue of St Andrew's cell subdivision, and then in the process of visualization, the algorithm can track the trend of wave surface to reduce cell computation and judgment. Finally, an example is given in which Couette movement reaches visualization of 2D-wave surface image. Source

Zhao H.,Hunan University | Zhao K.,Hunan University | Liu H.,Hunan University | Yu F.,Jiangsu Provincial Key Laboratory of Computer Information Processing Technology
Journal of Multimedia

Independent component analysis (ICA), instead of the traditional discrete cosine transform (DCT), is often used to project log Mel spectrum in robust speech feature extraction. The paper proposed using symmetric orthogonalization in ICA for projecting log Mel spectrum into a new feature space as a substitute in extracting speech features to solve the problem of cumulative error and unequal weights that deflation orthogonalization brings, so as to improve the robustness of speech recognition systems, and increase the efficiency of estimation at the same time. Furthermore, the paper studied the nonlinearities of the objective function in ICA and their coefficients, tested them in all kinds of environments, finding that they influenced the recognition rate greatly in speech recognition systems, and applied a new coefficient in the proposed method. Experiments based on HMM and Aurora-2 speech corpus suggested that the new method was superior to deflation- based ICA and MFCC. © 2012 ACADEMY PUBLISHER. Source

Zhang X.,Jiaxing University | Yue G.,Jiaxing University | Yue G.,Hunan University | Yue G.,Huaihua University | And 3 more authors.
Journal of Computers

In uncertainty decision making, experts can use interval-fuzzy number, triangular fuzzy number, trapezoidal fuzzy number, linguistic 2-tuple or linguistic indices to express their preferences. Using kernel function, the fuzzy numbers can be converted into fuzzy number range from 0 to 1. Thus, different fuzzy judgment matrix can be expressed in a unified style. A judging method for ordinal consistency of fuzzy judgment matrix was proposed according to the transitivity of binary relation. And two concepts of non-transitive route number(NTRN) and non-transitive route contribution number(NTCN) were put forward. Through the non-transitive route number and non-transitive route contribution number guidance, a revising method for fuzzy judgment matrix without ordinal consistency was put forward, in which the irrational element can be identified. The revising method can help the decision-maker revise his/her judgment matrix effectively. Finally, the non-financial performance evaluation attributes were selected and experts were asked to give the judgment matrix, fuzzy matrix without ordinal consistency was used to demonstrate the idea of the new judging and revising method for ordinal consistency of the fuzzy judgment matrix. © 2010 Academy Publisher. Source

Discover hidden collaborations