Entropy-Based Fabric Weave Pattern Indexing and Classification

Abstract

In textile design, fabric weave pattern indexing and searching require extensive manual operations. There has been little or no research on index and efficient search algorithms for fabric weave patterns. In this regard, we propose a method to index and search fabric weave patterns. The paper uses pattern clusters, boundary description code, neighbor transitions, Entropy and Fast Fourier Transform (FFT) directionality as a hybrid approach for the cognitive analysis of fabric texture. Then, we perform a comparison and classification of a wide variety of weave patterns. There are three common patterns used in textile design: (1) plain weave, (2) twill weave, and (3) satin weave. First, we classify weave patterns into these three categories according to the industrial weave pattern definition and weave point distribution characteristics. Second, we use FFT to describe the weave point distribution. Finally, an Entropy-based method is used to compute the weave point distribution and use this to generate a significant texture index value. Our experiments show that the proposed approach achieves the expected match for classifying and prioratizing weave texture patterns.