Non-Linear Extension of Generalized Hyperplane Approximation

Hyun-Chul CHOI  

IEICE TRANSACTIONS on Information and Systems   Vol.E99-D   No.6   pp.1707-1710
Publication Date: 2016/06/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.2015EDL8214
Type of Manuscript: LETTER
Category: Image Recognition, Computer Vision
non-linear regression,  feature augmentation,  generalized hyperplane approximation,  learning-based motion parameter estimation,  object tracking,  

Full Text: PDF>>
Buy this Article

A non-linear extension of generalized hyperplane approximation (GHA) method is introduced in this letter. Although GHA achieved a high-confidence result in motion parameter estimation by utilizing the supervised learning scheme in histogram of oriented gradient (HOG) feature space, it still has unstable convergence range because it approximates the non-linear function of regression from the feature space to the motion parameter space as a linear plane. To extend GHA into a non-linear regression for larger convergence range, we derive theoretical equations and verify this extension's effectiveness and efficiency over GHA by experimental results.