论文部分内容阅读
表情识别是近年来关于生物识别研究的热点课题之一,在心理学研究和计算机人性化方面具有深远意义。传统的通过面部特征和五官的相对位置进行表情识别方法由于受到光线和被测物表面复杂度的影响,并未有大的研究突破。本文根据表情的变化与面部温度分布的变化的关系提出了基于红外测温的表情识别方法,用通过分析面部温度区域分布的变化来分析表情的变化情况,用功能性分析的方法解决了传统方法困扰多年的问题。本文以人面部的红外热像图为研究对象,采用数学形态学方法,通过提取与分析热像图特征区域的整体几何特征和局部几何特征,进行了表情识别的研究。提出了温度直方图和基于鼻尖温度最低法的概念。本文中主要实现的方法有数学形态学的膨胀、腐蚀、开运算和闭运算,以及利用数学形态学手段进行二值图像的边缘检测。同时实现了面积、周长、似圆度和夹角的自动计算,作为分析特征区域的整体和局部几何特征的参数向量。通过数据分析,发现:随着表情变化,这些几何特征变化明显,并呈现一定的规律性,说明利用红外热像图研究表情识别是可行的。利用红外热像图进行表情识别,可以不受光线、肤色的影响,人脸区域和特征区域提取简单易行。本文主要利用了面部温度的分布变化进行了表情识别探索性研究,可以辅助心理学研究、高危病人检测和计算机人性化的发展等等。同时,根据温度分布变化同样可以进行人体生理功能的监测,辅助疾病的及时诊断和治疗。
Emoticon recognition is one of the hot topics in biometrics research in recent years and has far-reaching significance in psychology research and computer humanization. The traditional method of facial expression recognition based on the facial features and the relative positions of facial features does not make a big breakthrough due to the influence of the light and the surface complexity of the measured object. In this paper, based on the relationship between facial expression and facial temperature distribution, a facial expression recognition method based on infrared temperature measurement is proposed. The expression of facial expression is analyzed by analyzing the change of facial temperature distribution. The traditional method Troubled years of problems. In this paper, the infrared thermography of the human face is taken as the research object. The mathematical morphological method is adopted to study the facial expression recognition by extracting and analyzing the global geometric features and local geometric features of the thermal image features. The temperature histogram and the concept of the lowest tip temperature based method are put forward. In this paper, the main methods to achieve mathematical morphological expansion, corrosion, open computing and closed computing, as well as the use of mathematical morphology means binary image edge detection. At the same time, automatic calculation of area, perimeter, roundness and included angle is realized as a parameter vector for analyzing the global and local geometric features of the feature region. Through data analysis, it is found that with the change of facial expression, these geometric features change obviously and exhibit certain regularity, which shows that it is feasible to use infrared thermography to study facial expression recognition. The use of infrared thermography for facial expression recognition, can be unaffected by light, skin color, face area and feature extraction is simple and easy. In this paper, we make use of the change of facial temperature distribution to make an exploratory study of facial expression recognition, which can assist in the research of psychology, the detection of high-risk patients and the development of computerized humanity. At the same time, according to changes in temperature distribution can also monitor the physiological function of the human body, supporting the timely diagnosis and treatment of the disease.