基於背景模型的姿勢判斷系統

Arm Gesture Recognition Based on Background Model

李振遠(Chen Yuan Lee)

指導教授 ： 李忠謀

國立臺灣師範大學/理學院/資訊工程研究所/碩士(2011年)

摘要

姿勢辨識在電腦視覺領域中，特別是針對人體部分是項越來越重要的議題，涵蓋的範圍可分為：手部與手臂的姿勢辨識、頭部與臉部姿勢辨識、整個身體姿勢辨識等種類。在姿勢辨識的問題中，一個很大的瓶頸在於如何在複雜環境下取得所需要的特徵資訊，並且選擇適當的方法將這些資訊完成姿勢辨識。本論文主要目標是在真實的教室裡並且只有一台攝影機拍攝下，能即時(real-time)辨識出講者的手臂姿勢來達到控制投影片的效果，所提出的方法能讓講者在教室投影機照射下，穩定並不受投影機照射並且背景隨著投影片的換頁變化影響下抓取需要的資訊來進行辨識。本論文使用高斯混合背景模(Mixture of Gaussian background model)來擷取出前景(foreground)的輪廓(silhouette)影像，並使用連通元(connected component)將前景輪廓的特徵資訊截取出來，並套入支持向量機(Support Vector Machine，SVM)對手臂動作進行分類。此外，搭配人臉偵測(face detection)方法能分辨出左右手，達到不同手部動作來控制投影片的效果。

關鍵字

姿勢辨識 ； 高斯混合背景模型 ； 輪廓圖像 ； 連通元件 ； 支持向量機 ； 人臉偵測

並列摘要

Gesture recognition, recognize what poses a human body appears, has become an important issue in computer vision in recent. In general, gesture recognition considers different parts of human body, including head, hand and arm, and the whole body. In order to deal with gesture recognition, we need to well extract body silhouette even in a complex environment, to adopt features for gesture representation, and to design a proper classifier for recognition. In this thesis, our goal is to design a real-time presentation control system in a real classroom by recognizing the lecturer’s arm gestures only with single camera. Our proposed system is robust to strong lighting of projector and slide change in the projection screen. We first employ the mixture of Gaussian background model to segment the body silhouette of foreground. Then, the extracted feature of the body silhouette is classified as arm gestures by Support Vector Machine (SVM). In addition, the adaboosting approach of face detection helps our system to understand the left and the right hand to involve more hand actions for presentation control.

並列關鍵字

gesture recognition ； mixture of Gaussian background model ； silhouette image ； connected component ； Support Vector Machine ； face detection

參考文獻

[1] J. K. Aggarwal and Q. Cai, “Human motion analysis: A review,” Comput. Vis. Image Understanding, vol. 72, pp. 428–440, 1999.

[2] D. M. Gavrila, “The visual analysis of human movement: A survey,”Comput. Vis. Image Understanding, vol. 72, pp. 82–98, 1999.

[3] T. C. C. Henry, E. G. R. Janapriya, and L. C. deSilva, “An automatic system for multiple human tracking and actions recognition in office environment,” in Proc. ICASSP, 2003, vol. 3, pp. 45–48.

[4] J. Krumm, S. Harris, B. Meyers, B. Brumitt, M. Hale, and S. Shafer, “Multi-camera multi-person tracking for easy living,” in Proc. 3rd IEEE Int. Workshop Visual Surveillance, Jul. 2000, pp. 3–10.

[5] S. Dagtas, W. A. Khatib, A. Ghafoor, and R. L. Kashyap, “Models for motion-based video indexing and retrieval,” IEEE Trans. Image Process., vol. 9, no. 1, pp. 88–101, Jan. 2000.