Eyebrow semantic description via clustering based on Axiomatic Fuzzy Set

In this paper, we aim to extract the eyebrow semantic descriptors based on the Axiomatic Fuzzy Set (AFS) theory. First, we normalize the image of the eyebrows and automatically mark it by using a recently proposed facial landmarks detector. Second, a recent clustering algorithm based on AFS theory for eyes semantics abstraction is used to cluster these detected eyebrow landmarks and give semantic descriptors for each eyebrow. Finally, BU-4DFE and Multi-PIE databases are used to validate the effectiveness of the proposed approach. Furthermore, the eyebrow descriptions with different expressions and similar expressions are investigated and we show that the semantic descriptors are closely related to expressions. The experimental results show that the eyebrow semantic concepts obtained by the AFS clustering algorithm are better than the results produced by the traditional clustering methods (k-means and FCM) in terms of consistency for different expressions. This article is categorized under: Fundamental Concepts of Data and Knowledge > Knowledge Representation Algorithmic Development > Biological Data Mining