陈雅. 多媒体图像数据投影聚类融合算法优化研究[J]. 微电子学与计算机, 2017, 34(12): 134-137.
引用本文: 陈雅. 多媒体图像数据投影聚类融合算法优化研究[J]. 微电子学与计算机, 2017, 34(12): 134-137.
CHEN Ya. Multimedia Projection Image Data Fusion Algorithm Clustering Optimization Research[J]. Microelectronics & Computer, 2017, 34(12): 134-137.
Citation: CHEN Ya. Multimedia Projection Image Data Fusion Algorithm Clustering Optimization Research[J]. Microelectronics & Computer, 2017, 34(12): 134-137.

多媒体图像数据投影聚类融合算法优化研究

Multimedia Projection Image Data Fusion Algorithm Clustering Optimization Research

  • 摘要: 由于多媒体图像数据资源的种类繁多, 维数较大, 直接对其进行聚类融合的结果不理想, 需要对聚类融合算法进行优化.目前存在的基于邻域保持嵌入的多媒体图像数据投影聚类融合算法, 首先使用邻域保持嵌入法对多媒体图像数据投影样本维数进行约简处理; 然后对约简处理后的多媒体图像数据投影进行聚类融合.当前算法存在聚类性能较差, 融合效果不佳的问题.提出基于粒子群的多媒体图像数据投影聚类融合优化算法, 首先利用误差反传的梯度下降训练, 选取出多媒体图像数据投影的聚类成员, 为后续聚类融合提供准确的数据基础; 其次计算每个多媒体图像数据投影基聚类算法被选入优化基聚类子集的概率; 最后利用粒子群算法进行全局寻优, 实现对多媒体图像数据投影聚类融合算法的优化.通过实验验证分析, 结果表明, 所提算法可以提高多媒体图像数据投影融合质量和聚类准确率.

     

    Abstract: Because there are so many types of multimedia image data resources, dimension is bigger, directly on the clustering of fusion result is not ideal, need to optimize the clustering fusion algorithm. The existing based on neighborhood remain embedded multimedia projection clustering image data fusion algorithm, the first to use method of neighborhood remain embedded multimedia projection image data sample dimension reduction processing; Then the reduction process of multimedia image projection for clustering data fusion. The present clustering algorithm performance is poor, poor fusion effect.In this paper, a multimedia projection image data clustering based on particle swarm optimization algorithm fusion, first using the gradient descent of error back propagation training, selection of multimedia projection image data clustering members, provide accurate data foundation for the subsequent clustering fusion; A comprehensive evaluation of the definition of diversity and correctness and then use standard, realize the selectivity of multimedia projection image data clustering fusion; Finally on this basis using the particle swarm algorithm of global optimization for its optimization. Through the experiment analysis, the results show that the proposed algorithm can improve the quality of multimedia projection image data fusion and clustering accuracy.

     

/

返回文章
返回