A Novel Weave Pattern Encoding Method Using Neighbor Information and its Applications

Determination of features of weave patterns is a crucial task in practice, including determination of equivalent weave patterns, determination of weave type, repeat period detection, pattern correction, representation of fabric structures and pattern searching from a collection of weave patterns. Most previous research on the problems is based on binary matrix operations. In this paper, a simple and efficient encoding algorithm and its implementation are presented to reveal the hidden information in the binary matrix of a weave pattern and thus to find a solution to determine features of the weave pattern. By compiling the 32 components with different shapes of yarn cross-sections, it is possible to quickly produce the required weave geometries and weave textures at different levels of detail. A set of frequency matrices of a weave pattern can be considered as the fingerprint where each weave pattern is uniquely endowed with a set of frequency spectra. It therefore enables us to perform a weave pattern search in the database to locate weave geometries and weave textures. Our experiments report a satisfactory result on the problems.