研究主題 Researches

2016利用慣性量測單元與Kinect裝置發展上肢復健系統

五十肩與中風是現今常見的臨床疾病,這兩種疾病均會對患者身體造成活動不便,嚴重影響日常作息。因此如何有效提升復健療效與增加患者就診意願,將會是一項重要的課題。在醫療與科技的結合之下,虛擬體感系統隨之而來,其中虛擬體感的肢體動作辨識技術,更是被廣泛使用在復健醫學的研究領域中。由微軟所推出的Kinect視訊鏡頭,雖擁有低成本且開發方便等優點,但對於部分肢體的角度或是在肢體重疊的情況下,依然有較大的偵測誤判。本研究是將Kinect與無線傳輸的慣性感測器進行整合,設計出一套以上肢的復健動作為主的復健遊戲系統。利用Kinect 系統既有的骨架追蹤系統,進行遊戲的操作,同時在受測者的上肢穿戴多組慣性感測器,經由演算法計算各個上肢的角度,改善Kinect在深度計算上的誤差。受測時將完整的記錄兩套系統所計算出的肢體角度,並進行肢體角度的比較。研究結果顯示,Kinect在深度或是肢體重疊部分,Kinect會產生較大肢體角度錯誤,但透過IMU進行校正後,能有效改善Kinect在深度上的角度判斷,並能有更好的精確性與穩定性。希望未來能整合Kinect和慣性傳感器這兩項技術,藉由慣性傳感器進行Kinect的深度與肢體角度的修正,可以提升復健的療效與測量的信效度。

 

Frozen shoulder and stroke are both common clinical illnesses. Both illnesses may lead to disability and cause huge inconvenience on the patient’s daily routine. Therefore, improving rehabilitation efficacy and the increasing patient willingness to attend rehabilitation treatment are very crucial issues. Virtual Reality (VR) has been used in combination with medical treatment and technology. Its skeletal tracking system is being widely used in the field of medicine and rehabilitation. Although the Microsoft Kinect device with low cost and easy development, it may encounter the problem of misdetection when it comes to extremity angles or overlapping extremity. In this study, we proposed a new Rehabilitation Gaming System (RGS) which focuses on the upper part of the body with wireless Inertial Measurement Units (IMUs) and Kinect device. This system used Kinect as the base tracking system for its operation. Meanwhile, multiple sets of IMU were put on the extremity of the subject to calculate the angles through algorithms. The extremity angles calculated by the two systems during the test were recorded for further comparison. The results showed that Kinect has a greater error on extremity angles with regard to depth or overlapping extremities but calculation by IMUs has higher accuracy and stability. Therefore, we hope that Kinect and IMU could be integrated so that IMU can be used in assisting Kinect with depth and extremity angle correction in order to enhance the reliability and validity of rehabilitation efficacy and measurement.

 

          石正邦