arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1306.1913cs.CVcs.LGstat.ML
收藏

基于时间序列核的情感表情分类

Emotional Expression Classification using Time-Series Kernels

Eötvös Loránd University(罗兰大学);Carnegie Mellon University(卡内基梅隆大学);University of Pittsburgh(匹兹堡大学)

Andras Lorincz, Laszlo Jeni, Zoltan Szabo, Jeffrey Cohn, Takeo Kanade

摘要:

人脸表情的估计,作为时空过程,可以利用核方法,如果一个人考虑面部标志的位置和它们在3D空间中的运动。我们应用支持向量分类与内核来自动态时间规整相似性措施。我们仅使用标记点向量的PCA压缩表示的“运动模式”(即所谓的形状参数)实现了超过99%的准确性(通过ROC曲线下面积测量)。除了完整运动模式的分类之外,在从它们开始的5-6帧(大约200毫秒)内,以超过90%的准确度识别了若干表情。

英文摘要:

Estimation of facial expressions, as spatio-temporal processes, can take advantage of kernel methods if one considers facial landmark positions and their motion in 3D space. We applied support vector classification with kernels derived from dynamic time-warping similarity measures. We achieved over 99% accuracy - measured by area under ROC curve - using only the 'motion pattern' of the PCA compressed representation of the marker point vector, the so-called shape parameters. Beyond the classification of full motion patterns, several expressions were recognized with over 90% accuracy in as few as 5-6 frames from their onset, about 200 milliseconds.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑