A Novel Approach to Texture classification using statistical feature

TL;DR

This paper introduces a new texture classification method combining statistical features from Local Binary Pattern, Texture Spectrum, and Legendre Moments, achieving high accuracy for content-based image retrieval.

Contribution

The paper presents a novel combination of statistical features from LBP, Texture Spectrum, and Legendre Moments for improved texture classification.

Findings

Achieved 97.77% classification accuracy

Effective in content-based image retrieval

Combines multiple statistical features for robustness

Abstract

Texture is an important spatial feature which plays a vital role in content based image retrieval. The enormous growth of the internet and the wide use of digital data have increased the need for both efficient image database creation and retrieval procedure. This paper describes a new approach for texture classification by combining statistical texture features of Local Binary Pattern and Texture spectrum. Since most significant information of a texture often appears in the high frequency channels, the features are extracted by the computation of LBP and Texture Spectrum and Legendre Moments. Euclidean distance is used for similarity measurement. The experimental result shows that 97.77% classification accuracy is obtained by the proposed method.