5A3_0570_ KNITTED FABRIC TEXTILE CLASSIFICATION BASED ON GLOBAL FEATURE

Abstract

In this paper, we reviewed and examined the classification methods for the knitted fabrics and acquired a fabricdataset to verify the results. We then proposed the global feature method to classify 290 knitted fabrics with 58 different textures and 5 different colors. Using the global feature method, the classification accuracy of knitted fabrics can be improved significantly. The result shows that our proposed method is better than the stateof-the-art method.