Sparse Semi-Supervised Learning for Perceptual Grouping

Abstract

In this paper, we present a new perceptual grouping algorithm using sparse semi-supervised learning (SSSL). In SSSL, KD-tree is used for effective representation and efficient retrieval. SSSL performs both transductive and inductive inference with a new dynamic graph concept. The perceptual grouping problem is tackled using SSSL to group different patterns into one object and separate similar patterns into different objects. The proposed system is tested on three typical object patterns ranging from highly textured (zebra), to medium textured (tiger), to inhomogeneous appearance (horse). We compare the results with many alternatives such as normalized cuts, direct discriminative classification, conditional Markov random fields (CRF), and a discriminative structure learning algorithm. The overall results are promising, with several interesting empirical observations.

Cite

Text

Hong et al. "Sparse Semi-Supervised Learning for Perceptual Grouping." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2010. doi:10.1109/CVPRW.2010.5543640

Markdown

[Hong et al. "Sparse Semi-Supervised Learning for Perceptual Grouping." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2010.](https://mlanthology.org/cvprw/2010/hong2010cvprw-sparse/) doi:10.1109/CVPRW.2010.5543640

BibTeX

@inproceedings{hong2010cvprw-sparse,
  title     = {{Sparse Semi-Supervised Learning for Perceptual Grouping}},
  author    = {Hong, Yi and Jiang, Jiayan and Tu, Zhuowen},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
  year      = {2010},
  pages     = {1-8},
  doi       = {10.1109/CVPRW.2010.5543640},
  url       = {https://mlanthology.org/cvprw/2010/hong2010cvprw-sparse/}
}