Monogenic-LBP: A new approach for rotation invariant texture classification

$59

Description

We present a novel training free rotation invariant texture classification method, namely M-LBP. It combines two rotation invariant measures, the local phase and the local surface type extracted by the 1st- and 2nd-order Riesz transforms, with the traditional uniform LBP operator. 

Experimental results validate that M-LBP can achieve higher classification accuracy than the other methods evaluated, especially in the cases when the training set is small and not comprehensive.

Moreover, compared with the two state-of-the-art training based methods, MR8 and Joint, M-LBP has the advantage of smaller feature size and faster classification speed, which makes it a more suitable candidate in real applications.

 

 

ref :

Zhang, Lin, Lei Zhang, Zhenhua Guo, and David Zhang. “Monogenic-LBP: A new approach for rotation invariant texture classification.” In Image Processing (ICIP), 2010 17th IEEE International Conference on, pp. 2677-2680. IEEE, 2010.

Texture Segementation Algorithms for Electron Microscopy Images

Estimation and Preservation for Texture Enhanced Image Denoising

Reviews

There are no reviews yet.

Be the first to review “Monogenic-LBP: A new approach for rotation invariant texture classification”