A Hierarchical Probabilistic Model for Rapid Object Categorization in Natural Scenes
2011

A Model for Rapid Object Categorization in Natural Scenes

publication Evidence: high

Author Information

Author(s): He Xiaofu, Yang Zhiyong, Tsien Joe Z.

Primary Institution: Brain and Behavior Discovery Institute, Georgia Health Sciences University, Augusta, Georgia, United States of America

Hypothesis

How do humans achieve rapid scene categorization?

Conclusion

The model categorized animals and cars in natural scenes with near human-level performance.

Supporting Evidence

  • The model requires little training and is robust to variations in object categories.
  • It uses coarse hierarchical probability distributions to represent object categories.
  • The model integrates object localization and categorization.

Takeaway

This study created a model that helps computers quickly recognize objects in pictures, like animals and cars, just like humans do.

Methodology

The study developed a hierarchical probabilistic model that uses a small number of object structures for rapid categorization.

Limitations

The model does not account for temporal components or direct physiological correlates.

Digital Object Identifier (DOI)

10.1371/journal.pone.0020002

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication