漂移算法在视敏度检查中的应用

打开文本图片集

中图分类号：TP391.4文献标志码：A

Abstract： To achieve precise measurement of visual acuity of testers， a visual acuity detection system that combines probability distribution and mean shift algorithm has been developed. This system integrates the probability distribution with the mean shift algorithm， utilizing the characteristic of the mean shift algorithm， which is to explore the cluster centers along the direction of increasing data density， in conjunction with probability distribution theory， to accurately calculate the visual acuity probability corresponding to diffrent levels of visual acuity. Through repeated iterations， the algorithm continuously updates the exploration points， moving them towards areas of higher probability density until finding the point that most closely matches the actual level of the subject's visual acuity. In the implementation process， the algorithm adjusts the size of the visual markers in real-time based on the responses of the subjects.By continuously adjusting the observations towards their mathematical expectations or means， and through multiple rounds of iterations， the probability distribution gradually approximates the actual distribution of visual acuity. Experimental results show that， across three experimental scenarios， even in the case with the highest frequency of errors， the average error rate of the algorithm is only O.0o83. This essentially meets the requirements for stable and reliable visual acuity measurement， high robustness， and strong anti-interference capability.

Keywords： probability distribution; mathematical expectation; mean shift algorithm; visual acuity

随着信息科学技术和人工智能领域的迅速进步，眼视光学行业正在向智能化和数字化转型。（剩余10637字）