摘要
隨著電腦科技的進步，人機互動的改進一直是許多發表研究所專注的課題，從早期的鍵盤，滑鼠，到現在的攝影機，手寫筆、板等。都是為了能讓使用者能夠更方便地操作電腦。其演變趨勢從以機器為主，即使用者要學習如何操作電腦，轉變為以人為主，即由電腦配合人的習慣，讓使用者可以以自己的經驗習慣操作電腦。 本論文中提出了一套手勢辨識系統，可以在已知的環境底下去做手勢辨識，目的在於發展為一種新的人機互動機制，使用者僅須對攝影機比出已定義過的手勢，即可對電腦下命令。此方法不需要使用者戴手套或在指尖上塗顏色或安裝標誌即可辨識。其方法主要是以色彩資訊為基準，以HSI色彩空間為主，形態學的增長為輔找出手部區域及輪廓，並使用指尖模型找出指尖位置，利用所取得的資訊及連續影像的特性來預測指尖位置並追蹤之，最後經由追蹤到的結果來判斷使用者的手勢。
並列摘要
With the increasing advancing of pervasive computer and information technology, human-computer interaction has become a major research and development topic. The goal of this study is to involve human-computer interaction for ubiquitous interactions so as to provide a friendly service for users and to alleviate the burden of typing in keyboard and writing in tablet. In other words, the proposed method tries to recognize hand gesture, which in turn to provide a comfortable, attentive, and reliable system for human-computer interface. The proposed method consists of three phases: hand detection, fingertip tracing, and hand gesture recognition. In the stage of hand detection, skin-color information is used to detect the hand silhouette. In the stage of fingertip tracing, the edge detection and fingertip mask are involved to extract the fingertip trajectory. Finally, in the stage of hand gesture recognition, the fingertip trajectory vectors are extracted and matching strategy is adopted to achieve the goal of recognition. The effectiveness and practicability of the proposed method have been demonstrated by various experiments.