A Bag of Words approach for recognition of Greek folk dances
Eftychia Fotiadou, Ioannis Kapsouras, Nikos Nikolaidis, Anastasios Tefas · 2016
In this paper we explore the problem of distinguishing Greek folk dances from other kinds of activities, as well as from other dance genres, using video recordings. For this purpose, we adopt dense trajectories descriptors along with a Bag of Words (BoW) model to represent the motion depicted in the videos. Subsequently, a Support Vector Machine (SVM) classifier is used for classification. In order to evaluate the performance and the adequacy of the aforementioned framework for the dance recognition task, we performed experiments for three different classification setups: 1) Greek folk dances vs other activities (including different dance genres), 2) six different dance genres and 3) Greek vs Balkan folk dances. The aforementioned experiments exhibited satisfactory results, with successful recognition rates higher than 89.74%.