摘要
感兴趣区域(Region of interests,ROI)是图像中可能引起人眼视觉关注的区域。根据视觉注意机制的经典模型Itti模型来提取图像的低层特征,利用局部迭代的特征合并策略并在此基础上综合自动阈值分割和种子点的区域生长方法得到感兴趣区域的提取方法。实验结果表明该方法符合生物的视觉注意机制,具有良好的鲁棒性。
Region of interest is the area that may cause human visual attention.This article based on the classical model of visual attention Itti to extract low-level image features,using local characteristics of interation consolidation strategy and provides automatic thresholding and seeds based on region growing method of extraction of interest method.Experimental results show that this method is consistent with biological visual attention mechanism and has a good robustness.
出处
《电子设计工程》
2012年第7期160-162,共3页
Electronic Design Engineering
关键词
特征提取
感兴趣区域
显著图
视觉注意机制
feature extract
region of interest
saliency map
visual attention mechanism
