利用影像輪廓與臉形於人體方位偵測

Abstract

由於近年來智慧型機器人的發展迅速，機器人與人之間的互動也愈來愈頻繁，當機器人與人溝通時，必須知道人的位置、距離及面對的方向等資訊。在此篇論文中，我們結合二維影像輪廓比對與臉部方位偵測來完成三維人體方位偵測。首先，任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，進而獲得前景人物的輪廓。透過人體輪廓樣板比對及線性內插法，可以初步得到前景人物所朝的方向。當人臉朝向前方 30°以內時，也可透過雙眼與臉的三角幾何關係來估算人體所朝的方向。經實驗證明，我們提出的方法對於人體方位偵測的準確度相當高，平均角度誤差低於4°。

In recent years, the advancement of robot technology brings robot into human daily activities, and makes robot as an essential part of modern life. The interaction between human and robot has become more and more frequent. When robot interact with human, the information of the position, distance, and direction of a human becomes important. In the thesis, we combine 2D image silhouette matching and face direction detection to detect a human direction. Firstly, a foreground subject is extracted and converted to a binary image by a statistical background model. Then we obtain the silhouette of the foreground subject. Linear interpolation on silhouette matching results can be utilized to detect the human direction. When one’s face is within 30°, one’s pupils and their geometric relationship can also be exploited to estimate one’s direction. By numerical simulation, we have obtained a high accuracy, less than 4° on the average, in (or on) subject direction detection.