以前景物動態機率模型為基礎之嬰兒危險程度評估系統

蕭宛甄

指導教授 ： 方瓊瑤

國立臺灣師範大學/理學院/資訊工程研究所/碩士(2010年)

摘要

根據衛生署統計資料顯示，意外傷害是造成新生兒死亡的主要原因之一。意外傷害包含新生兒摔傷、被外物砸傷、被外物遮蔽口鼻窒息等。因此本論文希望開發一套視覺式的新生兒監控系統，可以即時的通知照顧者減少事故傷害的發生。
本論文主要分成三個部分：建立前景物顏色模型、嬰兒偵測、危險程度分析。監控攝影機架設於嬰兒床的正上方，系統讀入連續時間的新生兒動作影像後，先建立前景物顏色模型(FC model)。FC model建立時分別統計前景物的顏色及整張影像中所有的顏色，之後可作為嬰兒偵測時的依據。接著根據FC model，使用貝氏定理以顏色為特徵計算前景物的機率，另外根據MHI(motion history image)，以motion為特徵計算前景物的機率。最後顏色機率與motion機率結合，得到整合後的前景物機率完成嬰兒偵測。之後擷取出幾種代表性的特徵，代入函式得到個別的危險程度並依照新生兒月份做調整，最後整合所有特徵得到最終的危險程度值。本論文將危險程度分成五個等級，若超過系統設定的安全範圍則發出警告通知照顧者。

關鍵字

肢體動作 ； 監控 ； 嬰兒 ； 危險程度評估 ； 機率模型

並列摘要

This study presents a method for infant detection and degree of danger analysis. The system is divided into three parts: foreground color model (FC model) construction, infant detection, and degree of danger analysis. Before the infant detection, we construct FC model. In FC model construction, we calculate the foreground color histogram and the color histogram in an image, which are the basis for infant detection. Then we apply Bayes’ theorem to calculate the foreground probability based on color, addition. Furthermore, we according to MHI to calculate the foreground probability based on motion. Finally, we combine the color probability and the motion probability to obtain the foreground probability and to complete the infant detection. After that, we capture several features and calculate the individual degree of danger, and we adjust the degree of danger based on infant month age. Finally, we integrate all the features to obtain the final degree of danger. In this study, the degree of danger is divided into five levels. If it exceeds the system's safety range, the system will issue a warning to notice the baby-sitter.

參考文獻

[Hsu02] R. L. Hsu, M. Abdel-Mottaleb, and A. K. Jain, “Face detection in color images,” IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI 2002), Vol. 24, No. 5, pp.696-706, 2002.

[Zha08] Z. Zhang, H. Gunes, and M. Piccardi, “An accurate algorithm for head detection based on XYZ and HSV hair and skin color models,” Proceedings of the IEEE International Conference on Image Processing (ICIP 2008), San Diego, California, USA., pp. 1644-1647, 2008.

[Zhao08] M. Zhao, D. Sun, and H. He, “Hair-color modeling and head detection,” Proceedings of the World Congress on Intelligent Control and Automation (WCICA 2008), Chongqing, China, pp. 7773-7776, 2008.

[Hu04] W. Hu, T. Tan, L. Wang, and S. Maybank, “A survey on visual surveillance of object motion and behaviors,” IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews (SMC 2004), Vol. 34, No. 3, pp. 334-352, 2004.

[Gor07] L. Gorelick, M. Blank, E. Shechtman, M. Irani, and R. Basri, “Actions as space-time shapes,” IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI 2007), Vol. 29, No. 12, pp.2247-2253, 2007.