Colour in Context
Research group
Computer Vision Center

Who painted this painting?

CREATE - 2010
Download the publication : create_2010.pdf [525Ko]  
Significant research has been made in recent years in the field of automatic object and object category recognition. By describing local regions in images invariant to changes of viewpoint and illumination computers have shown to be able to reliably retrieve the same object from a data set of images. More recently, people have worked on category recognition, where task is to recognize an object class (e.g. cars or humans) based on a set of training examples. In this paper, we go one step further and investigate to what extend computers are able to infer the painter from a painting. We will use the bag-of-words approach which basically describes the statistics of small image patches. To test our approach we present a challenging data st of eight different painters. In our experiments color and shape features are used to classify paintings. The results obtained clearly demonstrate the significance of both color and shape features for painting classification.

Images and movies

 

BibTex references

@InProceedings\{SVV2010,
  author       = "Fahad Shahbaz Khan and Joost van de Weijer and Maria Vanrell",
  title        = "Who painted this painting?",
  booktitle    = "CREATE",
  year         = "2010",
  abstract     = "Significant research has been made in recent years in the field of automatic object and object category recognition. By describing local regions in images invariant to changes of viewpoint and illumination computers have shown to be able to reliably retrieve the same object from a data set of images. More recently, people have worked on category recognition, where task is to recognize an object class (e.g. cars or humans) based on a set of training examples. In this paper, we go one step further and investigate to what extend computers are able to infer the painter from a painting. We will use the bag-of-words approach which basically describes the statistics of small image patches. To test our approach we present a challenging data st of eight different painters. In our experiments color and shape features are used to classify paintings. The results obtained clearly demonstrate the significance of both color and shape features for painting classification.",
  url          = "http://www.cat.uab.cat/Public/Publications/2010/SVV2010"
}

Other publications in the database

 © 2008 Colour in context Group | Computer Vision Center. All rights reserved | Contact webmaster |  Last updated: Monday 11 May 2009     eXTReMe Tracker