基于人工智能的眼外肌外伤护理正位视训练系统设计  

Design of an artificial intelligence based orthostatic visual training system for external eye muscle injury nursing

作　　者：李丽兰 倪连红 连美燕 陈佳娜 LI Lilan;NI Lianhong;LIAN Meiyan;CHEN Jiana(The 910th Hospital of the Joint Logistics Support Force of the People's Liberation Army of China,Quanzhou Fujian 362100,China)

机构地区：[1]中国人民解放军联勤保障部队第910医院,福建泉州362100

出　　处：《自动化与仪器仪表》2023年第12期227-230,235,共5页Automation & Instrumentation

基　　金：2015年福建省自然科学基金项目(2015J01597)。

摘　　要：针对眼外肌外伤的正位视训练系统优化设计,研究设计了一种可穿戴式视线跟踪系统,并引入了基于图像梯度特征的眼球中心定位算法对其加以改进,同时将其应用到实际的视线跟踪中。结果显示,基于图像梯度特征的眼球中心定位算法在最坏误差为大于0.1的情况下,准确率均高于91.2%,且在大于0.2的情况下接近100%。同时利用视线跟踪系统测算的视线检测误差都比较低,且水平方向的检测误差最低仅为11.6,方差最低仅为7.3。说明该系统能够对改善眼外肌受损造成的眼位不正问题进行有效地正位视训练,同时为机器视觉领域的研究提供了可靠的技术参考。A wearable line of sight tracking system was studied and designed to optimize the ortho vision training system for extraocular muscle injury.An eye center localization algorithm based on image gradient features was introduced to improve it,and it was applied to practical line of sight tracking.The results show that the eye center localization algorithm based on image gradient features has an accuracy of over 91.2% in the worst-case error of greater than 0.1,and approaches 100% in the case of greater than 0.2.At the same time,the line of sight detection error calculated using the line of sight tracking system is relatively low,and the lowest detection error in the horizontal direction is only 11.6,and the lowest variance is only 7.3.This indicates that the system can effectively perform orthostatic training to improve the problem of eye misalignment caused by external eye muscle damage,and provides reliable technical reference for research in the field of machine vision.

关 键 词：眼球定位 视线跟踪 正位视训练 眼外肌外伤 人工智能 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置] R473.77[自动化与计算机技术—控制科学与工程]