-
[show abstract]
[hide abstract]
ABSTRACT: It is important to integrate contextual information in order to improve the performance of automatic image annotation. Graph based representations allow incorporation of such information. In this paper, we propose a graph-based approach to automatic image annotation which models both feature similarities and semantic relations in a single graph. The annotation quality is enhanced by introducing graph link weighting techniques based on inverse document frequent and the similarity of the word based on Co-occurrence relation in the training set . According to the characteristics of in ear correlation, block-wise and community-like structure in the modeled graph, we divide the graph into several sub graphs and approximate high rank adjacent matrix of the graph by using low rank matrix. Thus, we can achieve image annotation quickly. Experimental results on the Corel image dabasets show the effectiveness of the proposed approach in terms of performance.
Artificial Intelligence and Computational Intelligence (AICI), 2010 International Conference on; 11/2010
-
Sixth International Conference on Natural Computation, ICNC 2010, Yantai, Shandong, China, 10-12 August 2010; 01/2010
-
[show abstract]
[hide abstract]
ABSTRACT: In this paper, we summarize the characteristics of message search mechanism of Gnutella P2P network model. By analyzing its disadvantages, an improvement of resource discovery algorithm is proposed, it can more effectively reduce network traffic than the existing search methods in Gnutella system, and it can also improve Gnutella network's integrity, availability and scalability.
Natural Computation, 2007. ICNC 2007. Third International Conference on; 09/2007
-
[show abstract]
[hide abstract]
ABSTRACT: It is widely recognized that clustering ensemble is fit for any shape and any distribution dataset and that the boosting method provides superior results for classification problems. In the paper, a dual boosting is proposed for fuzzy clustering ensemble . At each boosting iteration, a new training set is created based on the original datasets' probability which is associated with the previous clustering. According to the dual boosting method, the new training subset contains not only the instances which is hard to cluster in previous stages , but also the instances which is easy to cluster. The final clustering solution is produced by using the clustering based on the co-association matrix. Experiments on both artificial and realworld datasets demonstrate the efficiency of the fuzzy clustering ensemble based on dual boosting in stability and accuracy.
Fuzzy Systems and Knowledge Discovery, 2007. FSKD 2007. Fourth International Conference on; 09/2007
-
Fourth International Conference on Fuzzy Systems and Knowledge Discovery, FSKD 2007, 24-27 August 2007, Haikou, Hainan, China, Proceedings, Volume 2; 01/2007
-
[show abstract]
[hide abstract]
ABSTRACT: In this paper, an effective synthetic aperture radar image segmentation method is proposed. Gaussian mixture models optimized by Greedy Expectation Maximization algorithm are applied. The immune genetic algorithm is employed to initialize Greedy Expectation Maximization algorithm and search the optimal values in the whole range, instead of general k-means algorithm, which is different from the traditional algorithm. Experimental results show our method can get better results for target segmentation. It can effectively segment the object from SAR images and inhibit speckle noise.
Artificial Intelligence and Computational Intelligence, International Conference on. 1:434-438.
