基於電腦視覺之手勢辨識系統

Abstract

在數位家庭的應用上，本篇論文提出一個方法來達到控制多媒體播放器的目的。此方法是以三維的人體模型為基礎，使用手勢辨識來提供一個直覺的控制方式。我們的動機是讓系統能夠察覺到使用者的意圖。為了追蹤使用者的手勢並且辨認這些手勢的含意以達到互動的控制系統的目的，貝式架構被運用在這個基於三維模型的手勢辨識系統中。另外，為了避免三維人體模型的高自由度可能造成手勢追蹤複雜化的問題，我們系統使用一個新的階層式演算法來增加系統效能。除此之外，這個系統應用多個特徵來增加追蹤的正確率。實驗結果可以看出本系統在於追蹤手勢是很穩定的，並且將來很有可能發展出一個控制多媒體的撥放器。

This paper presents a new approach to 3D model-based gesture recognition for controlling multimedia player. The motivation of this paper is to make home appliance aware of user’s intention. This 3D model-based gesture recognition system adopts a Bayesian framework to track the user’s hand posture and to recognize meaning of these postures for controlling 3D player interactively. To avoid the high dimensionality of the whole 3D upper body model, which may complicate the gesture tracking problem, our system applies a new hierarchical tracking algorithm to improve the system performance. Moreover, this system applies multiple cues for improving the accuracy of tracking results. Experimental results have shown that the proposed system tracks the hand posture robustly and has high application potential for controlling the multimedia player.