雙向臉部表情辨識應用於人與機器人互動

Bilateral Recognition of Facial Expressions for Human- Robot Interaction Applications

蔡明傑(Ming-Chieh Tsai)

指導教授 ： 羅仁權

國立臺灣大學/電機資訊學院/電機工程學研究所/碩士(2011年)

https://doi.org/10.6342/NTU.2011.10209

摘要

人口老化的現象引起社會福利醫療護理以及多種公共事業的需求。就技術上
而言，讓長輩接受智慧型的服務用機器人的幫助而有舒適、安全和健康的生活是一個非常重要的問題，所以智慧型機器人技術的領域將在21世紀會是一個高度優先發展的工業。
在機器人領域當中，機器人將採取於不同的臉部的表情辨識結果而產生一些行為。在表情辨識應用於人-機器人互動(HRI) 而言對於智慧型機器人技術領域中是有相當重要性的，所以我們對機器人怎樣與人臉部的表情辨識結果產生互動感興趣。例如:根據臉部的表情辨識，感情訊息被辨識出給機器人能更自然並聰明與人作互動是重要關鍵目的，讓機器人能理解人並且不單只透過滑鼠或者鍵盤所下達命令，因此我們發展一個即時臉部的表情辨識系統。在這篇論文裡，我們使用主動形狀模型得到特徵點，並加上人臉的色彩以結合提升準確度，用多支援向量機分類。支援向量機的核心選用倒傳遞基礎函數。根據分類結果，我們的機器人頭能根據臉部表情辨識達成互動，為臉部的表情辨識提出一個即時系統的一種方法，語音界面則採用微軟公司語音API(SAPI)來實現。
在這篇論文裡提議的全部系統，用戶界面，軟體框架和應用使用C++編程語言和Open CV，實驗在年輕的愛因斯坦機器人頭上，此智慧型的機器人技術和自動化被發展在國立台灣大學的IRA實驗室。

關鍵字

即時表情辨識 ； 情緒分類 ； 類人形機器人 ； 人-機器人介面

並列摘要

Growing of elderly population raises the demands of social welfare medical cares, and kinds of public services. In terms of technology, it is a very important issue to help elders live with a comfortable, safe and healthy life assisted by intelligent service robots. The field of intelligent robotics is a high-priority development industry in 21st century.
In the robotic area, robot will take some actions depending on different facial expressions. Human-Robot-Interaction (HRI) is an important role of intelligent robotics field, and we are interested in how robots interact with facial expressions. For example, according to facial expressions, robots can interaction with human. The recognition of emotional information is a key step toward giving robot the ability to interact more naturally and intelligently with human. Robot is able to understand humans and not just take orders via the mouse or the keyboard.
Therefore, we develop a real-time facial expressions system. In this thesis, we use active shape model to get feature points plus facial color to improve the accuracy and DAG support vector machine to classify them. Kernel of support vector machine is radial basis function. According to the recognition results, our Robot head can make facial expressions for human interaction. In this thesis, we propose a method which presents a real-time system for facial expressions. The speech interface is implemented with Microsoft Speech API (SAPI).
All the systems, user interface, software frameworks and applications proposed in this thesis are implemented with native C++ programming language and Open CV. The experiments are conducted on young Einstein robot head which is an intelligent robot developed by the Intelligent Robotics and Automation (IRA) Laboratory at National Taiwan University.

並列關鍵字

real-time facial expressions ； emotional classification ； Humanoid robot ； human-robot interaction.

參考文獻

[1] G. Yang and T. S. Huang, “Human Face Detection in a Complex Background,” Pattern Recognition, vol. 27, no. 1, pp. 53－63, 1994.

[3] T.K. Leung, M.C. Burl, and P. Perona, “Finding Faces in Cluttered Scenes Using Random Labeled Graph Matching,” Proceedings of 5th IEEE International Conference on Computer Vision, pp. 637－644, 1995.

[4] S. McKenna, S. Gong, and Y. Raja, “Modelling Facial Colour and Identity with Gaussian Mixtures,” Pattern Recognition, vol. 31, no. 12, pp. 1883－1892, 1998.

[5] M. Turk and A. Pentland, “Eigenfaces for Recognition,” Journal of Cognitive Neuroscience, vol. 3, no. 1, pp. 71－86, 1991.

[6] H. Schneiderman and T. Kanade, “Probabilistic Modeling of Local Appearance and Spatial Relationships for Object Recognition,” Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pp. 45－51, 1998