台灣物理治療師較缺乏採用電腦化評估系統來為病人做客觀的量測與評估之經驗。大部分的復健專業人士都是採用較不具準確性或有點較主觀的目測評估方式，來為病人完成評估量表的量測。因此，會很難客觀且準確地評估每次療程後的療效。本論文所提出一個電腦化物理治療評估系統，這個評估系統可以提供物理治療師有關病患的物理治療評估的準確資料。復健師基於系統所提供的評估資料，得以診斷出疼痛或行動失能的原因，然後，還能為每一位病患訂定其專屬的復健治療計畫。 基於不同的評估動作所需要的評估結果分為三套子系統，其中有平衡量測子系統、關節活動度量測子系統和協調度量測子系統。平衡量測子系統採用Wii Balance Board來測量靜態以及動態的站姿平衡，透過此子系統所提供的重心軌跡歷程資料和一些量化指標，物理治療師得以對病患進行平衡能力的評估。此外，又可透過此子系統中的平衡復健遊戲來增加病患進行復健的動機。至於另外的這兩個子系統，它們都有採用Kinect感測器來量測某些關節的活動度 (如：背、肩膀、手腕以及大拇指) 及協調能力 (如：手指到鼻子的評估動作量測)。針對不同的量測任務，本論文提出不同的影像處理演算法，來從Kinect所提供的景深資訊或身體輪廓資訊以完成特定之量測任務。目前所提出的三套子系統只有在實驗室的環境下進行測試實驗，未來將在醫院進行臨床測試以驗證系統的量測效果。 ;In Taiwan, physical therapists are usually lack of the experiences of adopting computerized evaluation systems to provide patients with objective measurements and assessments. Most of the rehabilitation professionals usually adopt the visual evaluations incorporated with a chart so that the assessment will be prone to be imprecise and subjective. Therefore, rehabilitation specialists encounter the problem of being unable to objectively assess the treatment effect after each individual course of treatment. This thesis proposes a computerized physical therapy evaluation system which is able to provide physical therapists with precise evaluations about patients’ physical conditions. Based on the evaluation outputs from the proposed evaluation system, rehabilitation professionals can diagnose the reasons of pains or movement dysfunctions and then provide patients with a customized rehabilitation exercise treatment program. The system consists of three sub-systems: the balance evaluation sub-system, the range of joint motion evaluation sub-system, and the coordination evaluation sub-system. The balance evaluation sub-system adopts a Wii Balance Board to measure the static and dynamic standing balances. It provides physical therapists with the information about the trajectories of the center of pressure and some quantitative indices about balance. In addition, a balance game is designed to promote patients’ interests in practicing the assigned rehabilitation exercises. As for the remaining two sub-systems, a Kinect sensor is adopted to measure the ranges of some joint motions (e.g., back, shoulder, wrist, and thumb) and the degree of coordination (e.g., finger to nose test). For each different evaluation assignment, a specific image processing algorithm was developed to process either the depth information or the skeletal tracking information to achieve the goal of evaluation. For the time being, the proposed evaluation system was tested in a laboratory setting. In the future, some pilot trials at hospitals will be carried out to test the performance of the system.
