Representing Texture Images using Asymmetric Gray Level Aura Matrices

• Author(s) / Creator(s)

  • Qin, Xuejie
  • Yang, Herb

• Technical report TR05-27. In this paper, we present a new mathematical framework for modeling texture images. Under this new framework, we prove that the Asymmetric Gray Level Aura Matrices (AGLAMs) of a given image have the necessary and sufficient information to represent the image. Using AGLAMs, a new similarity measure is defined, which is a one-to-one metric in the sense that zero distance between two images will guarantee that the two images are the same. To the best of our knowledge, none of the existing measures has this property. Applications such as learning for image retrieval and texture synthesis can be applied using AGLAMs. The experimental results show that the new AGLAM-based distance measure out-performs existing distance measures in the above mentioned applications. | TRID-ID TR05-27

• Date created

  2005

• Subjects / Keywords

  • Texture image retrieval
  • Texture synthesis
  • SVM
  • Texture similarity
  • AGLAM
  • Computer Graphics

• Type of Item

  Report

• DOI

  https://doi.org/10.7939/R31N7XM8G

• License

  Attribution 3.0 International