Conference/ProceedingsWorking Notes Proceedings of the MediaEval 2012 Workshop
Start date04.10.2012
End date05.10.2012
AddressSanta Croce in Fossabanda Piazza Santa Croce, 5 - 56125 - Pisa - Toscana - Italia
EditorMartha A. Larson, Sebastian Schmiedeke, Pascal Kelm, Adam Rae, Vasileios Mezaris, Tomas Piatrik, Mohammad Soleymani, Florian Metze, Gareth J.F. Jones
Author(s)Sebastian Schmiedeke, Pascal Kelm, Thomas Sikora
TitleTUB @ MediaEval 2012 Tagging Task: Feature Selection Methods for Bag-of-(visual)-Words Approaches
AbstractThis paper describes our participation in the Genre Tagging Task of MediaEval 2012, which aims to predict the videos’ category label. In last year’s participation, we performed experiments with bag-of-words (BoW) approaches in which different constellations in respect of modalities, features, and methods were investigated. This year, we focus on feature selection methods to improve the classification performance in terms of mean average precision (mAP) and classification accuracy (CA). We investigated the effectiveness of selection methods based on scores calculated using mutual informa- tion (MI) or term frequency (TF) and the effectiveness of transformation methods like the principle component anal- ysis (PCA).
Key wordsmultimedia analysis, multimedia application, genre classification, bag-of-words, SURF, Feature selection
NoteISSN 1613-0073