Texture features from Chaos Game Representation Images of Genomes

International Journal of Image Processing

Y. S. Thushana¹  Vrinda V Nair¹  Nisha N. S.¹  Vidya S.¹ 

[1] $$

关键词: Chaos Game Representation;    Texture analysis;    Wavelet decomposition;    Support Vector Machines;   

DOI  :  

来源: Computer Science Journals

【 摘 要 】

The proposed work investigates the effectiveness of coarse measures of the Chaos GameRepresentation (CGR) images in differentiating genomes of various organisms. Major work in thisarea is seen to focus on feature extraction using Frequency Chaos Game Representation(FCGR) matrices. Although it is biologically significant, FCGR matrix has an inherent error whichis associated with the insufficient computing as well as the screen resolutions. Hence the CGRimage is converted to a texture image and corresponding feature vectors extracted. Featuressuch as the texture properties and the subsequent wavelet coefficients of the texture image areused. Our work suggests that texture features characterize genomes well further; their waveletcoefficients yield better distinguishing capabilities.

【 授权许可】

Unknown