基于全局对比度的随机游走显著性视觉注意模型

Random Walk Saliency Detection Model Based on Global Contrast

模拟人类视觉的显著性视觉注意机制还没有形成统一的模型,依据对人类视觉的分析可知显眼、紧凑和对比度高的目标更加吸引人眼的注意,提出一种基于全局对比度结合随机游走的显著目标视觉注意算法,并将视觉显著性检测问题化为马尔科夫随机游走问题.首先计算输入图像的颜色和方向的全局对比度形成特征向量,利用向量间的距离确定图表示的边权重,从而构造随机游走模型的转移矩阵.同时通过全连通图随机游走和k_regular图随机游走提取图像的全局特性和局部特性,并将二者相结合得到显著图,从而确定显著目标.在国际上现有公开测试集上进行仿真实验,并与其它显著性视觉注意检测方法进行对比,结果表明,方法检测结果更加准确、合理,证明算法切实可行.

The saliency detection model simulating the human vision is not uniform.According to the analysis that a salient object in an image is often conspicuous, compact and of high contrast, a random walk saliency visual attention algorithm based on global contrast is pro- posed, converting the visual attention detection into markov random walk issue. First of all, the feature vector obtained by computing the color and orientation contrast values is used to determine the weight of edge, and then the isolated region is obtained by using the ran- dom walk on a complete graph to extract the global properties of the image. Meanwhile, the uniform region is enhanced by using the random walk on a k-regular graph to extract the local properties of the image. Finally, the saliency map is obtained by combining the global properties and local properties of the image, and the salient object is located and extracted according to the saliency map. Experimental results show that the proposed algorithm is more reasonable and effective than the other representative methods for salient object detection.