BOSSA: extended BoW formalism for image classification

Sandra Avila (1, 2) , Nicolas Thome (1) , Matthieu Cord (1) , Eduardo Valle (3) , Arnaldo de Albuquerque Araújo (2)

1 MALIRE - Machine Learning and Information Retrieval
2 NPDI - Lab. Nucleo de Proces.Digital de Imagens
3 RECOD - RECOD

In image classification, the most powerful statistical learning approaches are based on the Bag-of-Words paradigm. In this article, we propose an extension of this formalism. Considering the Bag-of-Features, dictionary coding and pooling steps, we propose to focus on the pooling step. Instead of using the classical sum or max pooling strategies, we introduced a density function-based pooling strategy. This flexible formalism allows us to better represent the links between dictionary codewords and local descriptors in the resulting image signature. We evaluate our approach in two very challenging tasks of video and image classification, involving very high level semantic categories with large and nuanced visual diversity.

Mots clés: SVM; Image classification; pattern recognition; Bag-of-Features; Bag-of-Words; visual dictionary; max pooling; sum pooling

Domaines: Apprentissage [cs.LG]; Traitement des images [eess.IV]

Sandra Avila, Nicolas Thome, Matthieu Cord, Eduardo Valle, Arnaldo de Albuquerque Araújo. BOSSA: extended BoW formalism for image classification. IEEE International Conference on Image Processing (ICIP), Sep 2011, Brussels, Belgium. pp.2909-2912, ⟨10.1109/ICIP.2011.6116268⟩. ⟨hal-00625533⟩