基于虚拟试验箱的康复手势识别算法研究

发布时间：2018-04-27 04:25

本文选题：虚拟康复 + 试验箱　；参考：《高技术通讯》2017年03期

【摘要】：研究了基于自主研发的手部康复训练系统的康复手势识别方法。针对现有手势识别算法识别手势过于单一、不具备针对性的问题,通过对手部功能障碍患者的手部运动及控制力的分析,提出了一种新的基于虚拟试验箱的康复手势识别算法。该算法的核心是利用摄像头捕捉不同的康复手势,并通过辅助训练标志板来实现手和辅助康复器械定位。基于改进的形状上下文识别算法的处理器实现了康复手势识别功能,并进一步控制虚拟场景中的物体做出相应的反应。该算法可以完成推、拉、悬垂、托举、二指捏等典型康复手势的准确识别,并与现有的手势识别算法进行了准确的对比。实验结果表明,该算法在识别率上有一定的提高,并且在识别的手势上更具有针对性。
[Abstract]:The method of hand hand gesture recognition based on hand rehabilitation training system is studied. Aiming at the problem that the existing hand gesture recognition algorithms are too simple to be targeted, a new rehabilitation gesture recognition algorithm based on virtual test box is proposed by analyzing the hand movement and control power of the patients with hand dysfunction. The core of the algorithm is to use camera to capture different rehabilitation gestures, and to achieve hand and rehabilitation device localization through auxiliary training marker board. The processor based on the improved shape context recognition algorithm realizes the rehabilitation gesture recognition function and further controls the objects in the virtual scene to react accordingly. The algorithm can be used to recognize typical rehabilitative gestures such as push, pull, drape, lifting, two-finger pinch and so on, and is compared with the existing gesture recognition algorithms. The experimental results show that the algorithm can improve the recognition rate and is more specific in the recognition of gestures.
【作者单位】：上海交通大学机械与动力工程学院;燕山大学电气工程学院;
【基金】：中国国家留学生基金资助项目
【分类号】：TP391.41

【参考文献】