Abstract Several state-of-the-art Generic Visual Categorization (GVC) systems are built around a vocabulary of visual terms and characterize images with one histogram of visual word counts. We propose a novel and practical approach to GVC based on a universal vocabulary, which describes the content of all the considered classes of images, and class vocabularies obtained through the adaptation of the universal vocabulary using class-specific data. An image is characterized by a set of histograms – one per class – where each histogram describes whether the image content is best modeled by the universal vocabulary or the corresponding class vocabulary. It is shown experimentally on three very different databases that this novel representation outperforms those approaches which characterize an image with a single histogram.
Adapted Vocabularies for Generic Visual Categorization
Computer Vision – ECCV 2006 ; 8 ; 464-475
Lecture Notes in Computer Science ; 3954 , 8
2006-01-01
12 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Adapted Vocabularies for Generic Visual Categorization
British Library Conference Proceedings | 2006
|Latent mixture vocabularies for object categorization and segmentation
British Library Online Contents | 2009
|Big Data and Controlled Vocabularies
NTRS | 2019
|Hybrid Generative-Discriminative Visual Categorization
British Library Online Contents | 2008
|Learning Vocabularies over a Fine Quantization
British Library Online Contents | 2013
|