Object Recognition

From Robot Intelligence
Jump to: navigation, search

Object Identification

  • Shape-Based Feature Point Extraction
    • Contour Fragment-based SPCF (Sampled Points from Contour Fragments) features
  • Descriptor Computation in Bisected Local Regions
    • Rotation-invariant shape contexts with geometric blur
    • Color histogram in the opponent color space
Iden blr.png
  • Spectral Matching
    • Pairwise geometric descriptors between features
Cate pairwise.png
  • Object Localization
    • RANSAC-based outlier removal


Object Categorization

  • Affine Category Shape Model
    • Contour fragment-based SPCF (Sampled Points from Contour Fragments) features
    • Rotation-invariant shape contexts with geometric blur
  • Spectral Matching
    • Pairwise geometric descriptors between features
  • Object Localization
    • RANSAC-based outlier removal
Cate overview.png
  • Results
Video


Related Papers

  • Yuli Chen, Yide Ma, Dong Hwan Kim, and Sung-Kee Park, "Object Recognition Based on a Simplified PCNN," in Proceedings of the 9th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2012), vol. 2, pp. 223-229, Rome, Italy, July 2012.
  • Dong Hwan Kim and Sung-Kee Park, "Object Class Detection Using Shape-Based Feature Descriptors," in Proceedings of the 7th International Conference on Ubiquitous Robots and Ambient Intelligence (URAI 2010), pp. 195-197, Busan, Korea, Nov. 2010.