結合時序姿態比對與模糊法則推論於人類動作辨識

Abstract

人類動作辨識在自動監視系統、人機界面、居家安全照護系統和智慧型居家環境等方面的應用中佔有主要的地位。許多人類動作辨識系統僅僅利用單一張影像的姿式來辨別該動作。但是，在時間序列上，姿式狀態轉換的關係是用來辨別人類動作的重要資訊。 在此篇論文中，我們結合時序姿態比對與模糊法則的方法來完成人類動作的識別。首先，每一張影像的前景人物利用一個基於前後影像比值而建立之統計背景模型抽取出來，並將抽取出來影像轉換成二值化的影像格式；此方法可以減少照明對前景人物抽取的影響。為達到較精準與可分別度，二值化影像經由特徵空間及標準空間轉換，投影至標準空間。最後人類動作的識別在標準空間中完成。經由樣板比對的方法可將三張影像序列，此影像序列乃從動作視訊5:1減低抽樣獲得，轉換成轉變成一組時序姿態序列。接著，利用模糊法則的推論方法，將這組時序姿態序列分類為某一個動作類別。模糊法則，不僅能夠結合時間序列上的資訊，並且可以容忍不同人做相同動作上的差異。在我們的實驗中，我們提出的動作辨認方法比利用HMM的方法，辨識正確率約增加5.4 %，達到91.8 %。

Human activity recognition plays an essential role in applications such as automatic surveillance systems, human-machine interface, home care system and smart home applications. Many of human activity recognition systems only used the posture of an image frame to classify an activity. But transitional relationships of postures embedded in the temporal sequence are important information for human activity recognition. In the thesis, we combine temple posture matching and fuzzy rule reasoning to recognize an action. Firstly, a foreground subject is extracted and converted to a binary image by a statistical background model based on frame ratio, which is robust to illumination changes. For better efficiency and separability, the binary image is then transformed to a new space by eigenspace and canonical space transformation, and recognition is done in canonical space. A three image frame sequence, 5:1 down sampling from the video, is converted to a posture sequence by template matching. The posture sequence is classified to an action by fuzzy rules inference. Fuzzy rule approach can not only combine temporal sequence information for recognition but also be tolerant to variation of action done by different people. In our experiment, the proposed activity recognition method has demonstrated higher recognition accuracy of 91.87% than the HMM approach by about 5.4 %.