摘要
视觉注意力能够对有限的信息资源进行分配,使感知具备选择能力。引入到图像分析领域,其模型的研究对自动估计图像的感兴趣区具有重要的意义。本文根据人类视觉感知的相关理论,对Itti视觉注意力模型进行了改进。首先,采用局部显著性度量的方法计算显著点的位置,然后融合进化规划和图像采样确定显著区域的大小,并根据注意焦点的转移依次得到一系列的显著区域。实验结果表明,用改进的注意力模型处理自然图像,获得了较为满意的效果。
Visual attention can distribute information resources appropriately, which makes human visual perception selective. In the domain of image analysis, researching on the visual attention model is significant, especially to extract the regions of interest. An improved Itti＇s visual attention model, inspired by human visual perception, is proposed. The salient point is located according to the local saliency measurement. And the size of the salient region is computed by combining evolutionary programming with subsampling. In this way, a series of salient regions are detected by shifting the focus of attention. The experimental results show that the method of the improved visual attention model is effective to process natural images.
出处
《电子测量技术》
2008年第3期1-3,10,共4页
Electronic Measurement Technology
基金
国家自然科学基金(60472036
60431020
60402036)
北京市自然科学基金(3052005)
教育部博士点基金(20040005015)
关键词
注意力模型
显著区域
注意焦点
visual attention model
salient region
focus of attention （FOA）
