度量对比敏感度阈值的自适应优化方法研究

Research on Adaptive Optimization Method for Measuring Contrast Sensitivity Threshold

对比敏感度作为视觉科学中人眼对物体分辨边界清晰程度的度量,表征了视觉对不同空间频率下对比度的响应能力,具有重要的研究意义.众多研究者给出了一系列用于解决度量对比敏感度阈值的方法,但现有方法在求解阈值时存在收敛速度慢、依从性差等缺点.本文提出了一种基于扰动算子的自适应psi-marginal算法.该算法在分析当前被试者反馈信息的基础上,通过扰动算子自适应优化刺激参数,使其快速向阈值附近逼近,提高收敛速度.实验结果表明,优化后的psi-marginal方法在保持阈值精度的基础上,平均收敛速度提升27.75%,其中当阈值在刺激范围的边缘时,收敛速度提升47.37%,改进方法效果较为明显.

Contrast sensitivity,as a measure of the clarity of the human eye′s distinguishing boundaries of objects in vision science,represents the response ability of vision to contrast at different spatial frequencies,and has important research significance.Many researchers have given a series of methods to measure the contrast sensitivity threshold,but the existing methods have the disadvantages of slow convergence and poor compliance when solving the threshold.This paper proposes an adaptive psi-marginal algorithm based on perturbation operator.Based on the analysis of the feedback information from the current subjects,the algorithm adaptively optimizes the stimulus parameters through the disturbance operator,so that it quickly approaches the threshold value and improves the convergence speed.Experimental results show that the optimized psi-marginal method can increase the average convergence speed by 27.75%while maintaining the accuracy of the threshold.Among them,when the threshold is at the edge of the stimulus range,the convergence speed increases by 47.37%,and the improved method is more effective.