一种基于视觉注意的小目标检测方法_谢文亮.pdfVIP

Computer Engineering and Applications 计算机工程与应用 2013 ，49 （12） 125 一种基于视觉注意的小目标检测方法 1，2 2 2 谢文亮 ，朱 丹 ，佟新鑫 1，2 2 2 XIE Wenliang , ZHU Dan , TONG Xinxin 1.中国科学院 研究生院，北京 100049 2.中国科学院 沈阳自动化研究所，沈阳 110016 1.Graduate University of Chinese Academy of Sciences, Beijing 100049, China 2.Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China XIE Wenliang, ZHU Dan, TONG Xinxin. Small target detection method based on visual attention. Computer Engineering and Applications, 2013, 49 （12）：125-128. Abstract ：A small target detection method based on visual attention is presented aiming at easily and quickly marking the targets in the image. In this method, Itti ’s common visual attention computational model has been improved and the target detection process is divided into bright target detection and dark target detection. In bright target detection, a simplified Itti ’s model is put to use, and in dark target detection, this paper mainly builds a lateral inhibition network model. It can synthetize the two saliency maps formed in bright target detection and dark target detection to get the final saliency map. After that a threshold is set to get the pre-attention region in the image. In order to mark each target relatively integral, fuzzy C-means algorithm is used for image segmentation. It shows that the proposed algorithm can detect the targets to be detected effectively. Key words ：visual attention; small target detection; saliency map; image segmentation 摘 要：为了方便而快速地在图像中标出目标，提出了一种基于视觉注意的小目标检测方法。该方法对Itti 通用视觉注 意计算模型作了改进，将目标检测过程分为亮目标检测和暗目标检测。其中亮目标检测采用了简化的Itti 模型，暗目标 检测主要采用侧抑制网络模型，将亮目标检测与暗目标检测各自生成的显著图合成得到最终的显著图。通过设定一阈 值得到图像的预注意区域，采用模糊C-均值算法进行图像分割，以便相对完整地标出每个目标。结果表明，算法能够有 效地检测待检测目标。 关键词：视觉注意；小目标检测；显著图；图像分割 文献标志码：A 中图分类号：TP391 doi ：10.3778/j.issn. 1002-8331.1110-0357 小目标检测