Document Image Classification Based on Improved Local Binary Patterns
Author(s): ZHANG Min
Pages: 827-831
Year: 2014 Issue: 10
Journal: Infrared Technology
Keyword: image classification; local binary pattern; texture analysis; dimensionality reduction;
Abstract: Based on the analysis of the methods to reduce the dimensions of the local binary pattern(LBP), a new operator called the orthogonal combination of local binary number(denoted as OC-LBN) is proposed for document image classification. Firstly, the local neighborhood is divided into different 4-orthogonal-neighbors, and the binary number of “1” in each 4-orthogonal-neighbor is used as its feature. Then, the features of all the 4-orthogonal-neighbor are combined together as region description. Experimental results obtained from texture, forward-Looking infrared and document image databases demonstrate that the proposed method can get the best performance of the methods mentioned in the paper.
Pages: 827-831
Year: 2014 Issue: 10
Journal: Infrared Technology
Keyword: image classification; local binary pattern; texture analysis; dimensionality reduction;
Abstract: Based on the analysis of the methods to reduce the dimensions of the local binary pattern(LBP), a new operator called the orthogonal combination of local binary number(denoted as OC-LBN) is proposed for document image classification. Firstly, the local neighborhood is divided into different 4-orthogonal-neighbors, and the binary number of “1” in each 4-orthogonal-neighbor is used as its feature. Then, the features of all the 4-orthogonal-neighbor are combined together as region description. Experimental results obtained from texture, forward-Looking infrared and document image databases demonstrate that the proposed method can get the best performance of the methods mentioned in the paper.
Related Articles