基于对比敏感度和马尔可夫链的注意信息提取算法

Extracting Attention Information Algorithm Based on Contrast Sensitivity and Markov Chain

借鉴生理学的研究成果,提出了一种新的基于对比敏感度和马尔可夫链的视觉注意信息提取算法.在注意特征向量提取之前,先用与离心率有关的对比敏感度函数对输入的图像进行加权,用以模拟视网膜神经节的反应机制；在特征向量上定义马尔可夫链,用它的平稳分布做为活动图上的显著度.算法的平均计算时间和以神经生物学家的研究成果为标准计算的接受者操作特性曲线下面积证明了算法的有效性.

Inspired by the research in physiology, a novel algorithm for extracting bottom-up attention information （integra- tion of contrast sensitivity and Markov chain, ACSMC） is proposed in this paper. In our algorithm, the original image is weighted with a contrast sensitivity formula which is a ftmction retinal eccentricity to simulate the mechanism of retinal ganglion. A Markov chain is defined on feature maps. The equilibrium distribution of this chain is taken as saliency values. The average of algorithm cost time and area under receiver operating characteristic curve （AUROC） based on the research of neurobiologist demonstrate its effec- tiveness.