New semantic descriptor construction for facial expression recognition based on axiomatic fuzzy set
New semantic descriptor construction for facial expression recognition based on axiomatic fuzzy set
复制标题
DOI：
10.1007/s11042-017-4818-3
复制
发表时间：
2018-05
影响因子：
3.6
通讯作者：
Zedong Li;Qingling Zhang;X. Duan;Cun-rui Wang;Yu Shi
中科院分区：
文献类型：
--
作者：
Zedong Li;Qingling Zhang;X. Duan;Cun-rui Wang;Yu Shi
本文提出了一种基于公理模糊集（AFS）的人脸表情语义描述器。新的描述符有两个优点：第一，它不依赖于优先知识，当一个使用它来构建语义概念。根据特征数据的分布情况，利用模糊隶属度可以快速建立语义概念。二是描述子通过对语义概念的操作来描述复杂的特征。所开发的描述符可以提供表情特征的变化和关系。最后，我们在FEI和CK+数据库上实现了我们的方法，并对各种表达式进行了语义解释。同时，采用C4.5、贝叶斯、决策表、Cart和减少错误剪枝树等方法对算法的性能进行了评估。
In this paper, we propose a new semantic descriptor based on axiomatic fuzzy set (AFS) to describe facial expressions. The new descriptor has two advantages: The first one is that it does not depend on priori-knowledge, when one uses it to construct semantic concepts. According to the distribution of feature data, one can quickly establish semantic concepts using the fuzzy membership degree. The second one is that the descriptor can describe complex features by implementing operation on semantic concepts. The developed descriptor can provide variations and relations of expression features. Finally, we implement our method on FEI and CK+ database, and make semantic interpretations for various expressions. Meanwhile, the performance is evaluated with the state-of-the-art methods such as C4.5, Bayes, Decision Table, Cart and Reduced error pruning tree.