Classification of Natural Scenes

@inproceedings{DrewesClassificationON,
  title={Classification of Natural Scenes},
  author={Jan Drewes and Dekan Prof and J{\"u}rgen Hennig and Gie\ssen Psychologie and Karl Berichterstatter and Gegenfurtner and Frank Bremmer}
}
Humans are capable of rapidly classifying scenes by content, even when they are presented only very briefly. Classification accuracy can exceed 90%, while above-chance performance can be achieved in about 150ms. The global amplitude spectrum of an image has repeatedly been suggested to be a possible source of information for such fast classification. The aim of this thesis was to analyze the way in which humans classify images, specifically for the case of scenes which contain an animal or not… CONTINUE READING

References

Publications referenced by this paper.
Showing 1-10 of 40 references

The time course of visual processing: Backward masking and natural scene categorisation

  • Bacon-Mace
  • Vision Research
  • 2005

A, Statistics of natural image categories, Network: Computation in Neural Systems

  • Oliva Torralba, A Torralba, Oliva
  • A, Statistics of natural image categories…
  • 2003

Illustration 12: Spectral signatures of different image categories (taken from [Torralba

  • Illustration 12: Spectral signatures of different…
  • 2003

Mean amplitude spectra of " Animal " and " Nature " image (Taken from: [Johnson

  • Illustration
  • 2003

Similar Papers

Loading similar papers…