ID 原文 译文
8314 针对当前基于进化算法的相关反馈图像检索方法无法很好地结合用户偏好信息和设置参数过多的问题,提出一种基于改进教与学优化的相关反馈图像检索方法。 In view of the current relevance feedback image retrieval method based on evolutionary algorithms cannot be well combined with the user preference information and set parameters too much problem, put forward an improved optimization of teaching and learning based on relevance feedback image retrieval method.
8315 根据图像检索问题的特定环境,对教与学优化算法进行了一系列改进。 According to the specific environment of image retrieval problem, the optimization algorithm in a series of teaching and learning to improve.
8316 首先,结合最近邻分类法构造适应度函数的约束条件,使之更好地反映用户偏好信息; first of all, combined with the nearest neighbor classification constraints of fitness function is constructed, to better reflect the user preference information;
8317 其次,通过在教阶段将相关图像集的中心图像作为教师以及在学阶段将相关图像作为学员学习的对象,使算法快速收敛到相关图像区域; Secondly, through the related teaching stage image set at the center of the image as a teacher and learning phase the related image as an object of students learning algorithm fast convergence to the relevant image area;
8318 最后,结合约束处理技术Deb准则进行学员的选择操作。 Finally, combined with the constraint processing Deb criterion for the choice of students.
8319 将该算法与目前效果优异的3种基于进化算法的相关反馈技术在两套标准图像测试集上进行对比。 Would the algorithm and the good effect of three kinds of relevance feedback technique based on evolutionary algorithms in the two sets of standard images were compared on the test set.
8320 结果表明,所提算法相较于另外3种算法具有明显的优势,能更好地结合用户偏好信息提高图像检索性能。 The results show that the proposed algorithm compared with other three kinds of algorithm has obvious advantages, and better able to combine the user preference information improve image retrieval performance.
8321 针对分布式多输入多输出(multiple input multiple output,MIMO)雷达在多任务模式下对目标快速跟踪的需求,结合"低慢小"目标的特点,提出了基于多目标位置跟踪的收发阵元选取算法。 For distributed multiple input multiple output (multiple input multiple output, MIMO) radar in multitask mode to target the needs of the fast track, combining with the characteristics of "low slow small target, based on multi-objective location tracking is proposed to send and receive array yuan selection algorithm.
8322 首先,以选取最小的阵元集合为代价函数,在指定的位置估计精度的约束下,建立了阵元选取的优化模型。 First, in order to select the smallest array yuan set as the cost function, estimation precision in the location specified, under the constraint of array element selection optimization model is established.
8323 然后,将该优化问题看作成背包问题,提出了改进的阵元选取算法对模型进行求解。 Then, the optimization problem as knapsack problem, array selection algorithm was put forward to solve the model.