A Study on the Extraction and Application of Image Contour Feature Codes

Yongfen Wu, Wei Xu, Daoqiang ZhangJournal of Residuals Science and Technology 2016

Abstract

To extract image contour feature codes more effectively, this paper reprocesses the acquired boundary contour information and extracts more effective image contour feature codes. The innovation of this algorithm is that based on contour coordinates extracted by the level set evolution algorithm, after a series of conversions, multiple two-dimensional contour matrixes of the same size can be obtained. After that, matrixes are coded horizontally, column-wise and diagonally, to get new coded features. Through an anti-interference analysis of the algorithm, results show that the proposed feature code extraction algorithm in this paper has high stability and the extraction schemes are flexible and variable. It is easier to acquire distinguishable information in very similar images. To verify the validity of this algorithm, the author applies the feature code exaction algorithm to the field of facial expression recognition and establishes a facial expression recognition model based on facial part contour code exaction. Experimental results show that the facial expression recognition system established in this paper can overcome the recognition interference caused by strong similarity between samples. The comprehensive facial expression recognition rate reaches more than 97.20%.
BibTeX
@article{wu2016study,
  author = {Yongfen Wu and Wei Xu and Daoqiang Zhang},
  title = {{A Study on the Extraction and Application of Image Contour Feature Codes}},
  journal = {Journal of Residuals Science and Technology},
  year = {2016},
}

Loading related content...

Figure