Entity

Time filter

Source Type


Tang P.-S.,University of Electronic Science and Technology of China | Tang X.-L.,University of Electronic Science and Technology of China | Tao Z.-Y.,University of Electronic Science and Technology of China | Li J.-P.,University of Electronic Science and Technology of China | Li J.-P.,Chongqing Key Laboratory of Computer Network and Communication Technology
2014 11th International Computer Conference on Wavelet Active Media Technology and Information Processing, ICCWAMTIP 2014 | Year: 2014

The wide application of Internet technology and media technology produces more and more data which also leads the arrival of the era of big data. However, it is difficult to extract the needed information from the original data directly except some special conditions. In recent years, the development of machine learning which provide a effective way to solve this problem for us. You can obtain lower rate of Miscalculate when you select a reasonable feature selection algorithm under the premise of not increasing the complexity of algorithm. At present it is divided into two categories named the Filter and Wrapper feature selection algorithm in the field of machine learning. This paper considers both the advantages and disadvantages of these two feature selection algorithm and studies the combined feature selection algorithm. © 2014 IEEE. Source


Chen J.,University of Electronic Science and Technology of China | Li J.,University of Electronic Science and Technology of China | Li J.,Chongqing Key Laboratory of Computer Network and Communication Technology | Lin J.,University of Electronic Science and Technology of China | Huang Y.,Chengdu University of Information Technology
2014 11th International Computer Conference on Wavelet Active Media Technology and Information Processing, ICCWAMTIP 2014 | Year: 2014

In this paper, we utilize negacyclic codes to construct a family of optimal subsystem codes with parameters [[q2+1,(q-1)2, 4, q-1]]q, where q ≡ 1(mod4), q = pτ, τ ≥ 1 and p is an odd prime. These constructed quantum subsystem codes are different from those codes available in the literature. © 2014 IEEE. Source


Tang X.,University of Science and Technology of China | Tao Z.,University of Science and Technology of China | Tang P.,University of Science and Technology of China | Li J.,University of Science and Technology of China | Li J.,Chongqing Key Laboratory of Computer Network and Communication Technology
2014 11th International Computer Conference on Wavelet Active Media Technology and Information Processing, ICCWAMTIP 2014 | Year: 2014

Pulse-Coupled Neural Network is known as third generation artificial neural network. It is created by visual cortex neurons, a synchronous pulse release phenomenon of mammals. Compare to traditional artificial neural network, PCNN has the characteristics of dynamic neural network, integrated space-time, automatic propagation and synchronous pulse release. PCNN has tendency for image retention information and edge detection. © 2014 IEEE. Source


Huang Y.,University of Electronic Science and Technology of China | Huang Y.,Chongqing Key Laboratory of Computer Network and Communication Technology | Huang Y.,Chongqing University of Posts and Telecommunications | Xie M.,University of Electronic Science and Technology of China | And 2 more authors.
Proceedings - 2011 8th International Conference on Fuzzy Systems and Knowledge Discovery, FSKD 2011 | Year: 2011

Fractional differential is used to adjust the affinity matrix of the image and to extract more useful edge and texture information. A graph is constructed for the original image with edge and texture information calculated with fractional differential, and the Laplacian matrix is obtained and the eigenvector corresponding to the second eigenvalue is classified to get the final segmentation results. Experimental results show that the proposed method can extract more fine details, preserve more useful information and get more accurate results. © 2011 IEEE. Source


Huang Y.,University of Electronic Science and Technology of China | Huang Y.,Chongqing Key Laboratory of Computer Network and Communication Technology | Huang Y.,Chongqing University of Posts and Telecommunications | Xie M.,University of Electronic Science and Technology of China | And 3 more authors.
Journal of Information and Computational Science | Year: 2012

Algebraic Multi-grid (AMG) method is analyzed and is combined with the graph cut method to extract the texture feature of the image. AMG method can extract more information about the singularities in the image. The data in the coarse grid is utilized to reconstruct the image and the reconstruct result approximates to the original image well. On the basis of the analysis on the AMG method, an energy function is constructed for the texture feature and minimized using max-flow method. Experimental results show that the proposed method can extract more texture details accurately. © 2012 by Binary Information Press. Source

Discover hidden collaborations