基于虚拟试验箱的康复手势识别算法研究

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。

Research on an algorithm for hand gesture recognition based on virtual experiment box

The hand gesture recognition method based on an independently developed hand rehabilitation training system was studied, and a new rehabilitation gesture recognition algorithm based on virtual test box was proposed through the analysis of the hand movement and control of the patients with hand dysfunction to overcome existing algorithms'' problems of too simple and without pertinence.The core of the algorithm is using a camera to capture different gestures and achieve the location of hands and rehabilitation equipment with an auxiliary marker board.The processor of the algorithm based on improved shape context recognition can achieve the rehabilitation gesture recognition, and control the virtual scene in the object to make the corresponding response.The algorithm can currently distinguish the gestures of push, pull, drape, lifting and two finger pinch and compare the recognition with the existing algorithms.The experiment results show that the proposed algorithm can improve the recognition rate and is more targeted in gesture recognition.