基于特定区域纹理描述和面部动态变化的自发性微笑判别技术（英文）

发布时间：2018-04-16 22:36

本文选题：面部特征定位 + 几何特征　；参考：《Frontiers of Information Technology & Electronic Engineering》2017年07期

【摘要】：微笑作为一种典型的生物多样性特征信号,在社会交往中有较大影响力,它揭示了人的情感感受和内心状态。自发性的微笑与假笑由不同大脑系统发出,在形态学和动力学上均存在差异。区分这两种类型的微笑仍具有挑战性,因为其中细微差别很难被肉眼观察到,仍有待被识别捕捉。已有相关研究大多是提取自发性微笑的几何特征,而外观特征并没有被充分利用,导致纹理信息的丢失。本文提出一种基于特定区域纹理描述来表示不同面部区域的局部模式变化,从而弥补几何特征研究的局限性。每个面部区域的时间相位是通过计算相应的面部区域强度来划分,而非仅考虑嘴巴区域强度。同时利用支持向量机的中层融合策略,将两种特征类型结合起来。实验结果表明,本文提出的外观表示法及其与基于几何形状的人脸动力学的结合技术,在BBC、SPOS、MMI和UvA NEMO四个基准数据库中得到很好的应用。
[Abstract]:As a typical signal of biodiversity, smile has great influence in social communication. It reveals the emotional feelings and inner state of people.Spontaneous smiles and fake smiles are produced by different brain systems, and there are differences in morphology and dynamics.Distinguishing between the two types of smiles is challenging because the nuances are hard to see and yet to be recognized and captured.Most of the related studies have extracted the geometric features of spontaneous smile, but the appearance features have not been fully utilized, resulting in the loss of texture information.In this paper, a texture description based on specific regions is proposed to represent the local pattern changes of different facial regions, which can make up for the limitations of geometric features.The temporal phase of each facial region is divided by calculating the corresponding strength of the facial region, rather than considering the strength of the mouth region.At the same time, the support vector machine (SVM) middle level fusion strategy is used to combine the two feature types.The experimental results show that the appearance representation presented in this paper and its combination with geometric shape based face dynamics are well applied in the four reference databases of BBCs SPOS MMI and UvA NEMO.
【作者单位】： MOE
【基金】：supported by the National Natural Science Foundation of China(No.60675025) the National High-Tech R&D Program(863)of China(No.2006AA04Z247) the Scientific and Technical Innovation Commission of Shenzhen Municipality,China(Nos.JCYJ20130331144631730 and JCYJ20130331144716089) the Specialized Research Fund for the Doctoral Program of Higher Education,China(No.20130001110011)
【分类号】：TP391.41

【参考文献】