Detecting Fabric Pattern Infringement Depending On Content-Based Image Retrieval Technique

With the improvement of people’s living standards,the demand for personalized fabric product is more and more intense.And the fabric pattern is more and more valuable which leads to the growing number of fabric pattern infringement cases.This paper proposes and implements a fabric pattern infringement detection system that can quickly and accurately detect infringement, thereby protecting the interests of the fabric pattern copyright holders.The system is implemented depending on content-based image retrieval technique,which consists of two modules,that is online module and offline module.To make the detection more accurate,we take ASIFT and texture feature as the image feature descriptors according to the application characteristics that the infringing pattern is often through translation,rotation,scale transformation,viewpoint change and shelter processing, and the fabric pattern is with notable texture.We have proposed a method called rectangle expansion to deal with the ASIFT feature redundancy which is derived from repeatability in fabric pattern. To overcome the deficiencies of traditional method,we propose a new ASIFT feature similarity formula.Finally,we conduct experiments to determine the parameters of rectangle expansion method and test the effectiveness of the rectangle expansion method and the fabric pattern infringement detection system.