Effective Use of Frequent Itemset Mining for Image Classication

Basura Fernando (1) , Elisa Fromont (2) , Tinne Tuytelaars (1)

1 ESAT - Department of Electrical Engineering [Leuven]
2 LHC - Laboratoire Hubert Curien

In this paper we propose a new and e ective scheme for applying frequent itemset mining to image classi cation tasks. We refer to the new set of obtained patterns as Frequent Local Histograms or FLHs. During the construction of the FLHs, we pay special attention to keep all the local histogram information during the mining process and to select the most relevant reduced set of FLH patterns for classi cation. The careful choice of the visual primitives and some proposed extensions to exploit other visual cues allow us to build better bag-of-FLH-based image representations. We show that these bag-of-FLHs are more discriminative than traditional bag-of-words and yields state-of-the art results on various image classi cation benchmarks.

Mots clés: Frequent patter; image classification; discriminative pattern mining

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]

Basura Fernando, Elisa Fromont, Tinne Tuytelaars. Effective Use of Frequent Itemset Mining for Image Classication. European Conference on Computer Vision, Oct 2012, Firenze, Italy. pp.214-227. ⟨hal-00714712v2⟩