数字化图像动态目标搜索眼动特征增强仿真

Enhanced Simulation of Eye Movement Feature for Dynamic Target Search in Digital Image

针对已有的数字化图像目标搜索算法忽略了对提取特征的增强处理,导致传统算法存在平均查准率低、虚警率高的问题,提出数字化图像目标搜索眼动特征增强算法。通过热点图方法获取眼动特征,并利用量子遗传算法对获取的眼动特征筛选和提取,实现眼动特征增强。将增强后的眼动特征输入到支持向量机的决策函数中,实现数字化图像目标搜索。仿真结果表明,所提算法的平均查准率高、虚警率低、目标搜索概率高、搜索效率理想。

Some existing algorithms often ignore the enhancement for extracted features, resulting in low average precision and a high false alarm rate during the digital image target search. Therefore, an algorithm for enhancing eye movement characteristics in digital image target search was proposed. The eye movement characteristics were obtained by a hot spot map. And quantum genetic algorithm was used to screen and extract the obtained eye movement characteristics, thus enhancing these characteristics. Moreover, the enhanced characteristics were input into the decision function of the support vector machine. Finally, the digital image target search was achieved. Simulation results show that the proposed algorithm has high average precision, low false alarm rate, high target searching probability, and ideal search efficiency.