Statistical Model for Occluded Object Recognition

  • Authors:

  • David Castanon

  • Affiliations:

  • -

  • Venue:
  • ICIIS '99 Proceedings of the 1999 International Conference on Information Intelligence and Systems
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we present a model-based statistical algorithm for recognition of partially occluded objects from noisy features. The likelihood ratio of the image features to template features is used for recognition. Two different statistical occlusion models are introduced: an independent prior model and a Markov Random Field (MRF) prior model. Our experiments show that the MRF model performs more robustly than the independent model in the presence of partial occlusion.