多媒体图像数据投影聚类融合算法优化研究

发布时间：2018-03-26 00:23

本文选题：多媒体图像数据投影　切入点：聚类融合算法　出处：《微电子学与计算机》2017年12期

【摘要】：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]:Based on particle swarm optimization (PSO), an optimal clustering algorithm for multimedia image data projection is proposed. Firstly, clustering members of multimedia image data projection are selected by using gradient descent training of error backpropagation, which provides an accurate data basis for subsequent clustering fusion. Secondly, the probability of each multimedia image data projection clustering algorithm being selected into the optimal base clustering subset is calculated. Finally, the particle swarm optimization algorithm is used for global optimization. The experimental results show that the proposed algorithm can improve the fusion quality and clustering accuracy of multimedia image data projection.
【作者单位】：淮阴师范学院计算机科学与技术学院;
【分类号】：TP18;TP311.13

【相似文献】