Research on Mind Analysis System Robust under Environmental Variation

本研究的目的是提出一种人脸表情分析方法，并建立人脸表情分析系统。我们使用摄像机拍摄的人脸图像，并对图像应用了几种图像处理方法，该分析方法对环境变化，如光照变化，具有较强的鲁棒性。2005年，我们考虑了脸型测量方法，建立了测量系统。2007年，我们考虑了对光照变化具有较强稳健性的嘴巴跟踪方法，并编制了分析程序。然后，我们使用在光照强度变化中拍摄的人脸图像对该程序进行了实验，并证实了对人脸区域的稳健跟踪。此外，我们认为该方法对其他几种变化，如光照位置变化和光照角度变化都是稳健的。我们提出的方法可以在几个真实环境下的几个应用中使用。

The purpose of our study is to propose the facial expression analysis method and to create the analysis system. We use facial images taken by a video camera and apply several image processing methods to the images, and this analysis method is robust for the environmental variation, such as the lighting variation. In the year of 2005, we have considered the facial shapes measuring method and created the measuring system. In the year of 2007, we have considered the mouth tracking method, which is robust for the lighting variation, and we have created the analysis program. Then, we have experimented with the program using the facial images that are taken in the lighting intensity variation, and we have confirmed the robust tracking of the facial regions. Furthermore, we have considered the method robust for several other variations, such as lighting position variation and lighting angle variation. Our proposed method can be used in several applications under several real environments.