Printed Image Segmentation Based on Fast Fuzzy Clustering of Super-Pixels
Aiming at the current problems of low efficiency and high complexity in the color difference detection process of color prints,a print image segmentation method(SFFCM)based on fast fuzzy clustering of super-pixels is proposed.A simple linear iterative clustering(SLIC)algorithm is first used to segment the image into closely neighboring super-pixel regions.Each super-pixel region is considered as an independent clustering unit.Subsequently,the fuzzy C-mean clustering(FCM)algorithm is applied to the computation of the attribution relationship of the super pixels,i.e.,the value of affiliation is introduced to allow the super-pixels to belong to more than one clustering centers and fuzzy clustering is achieved by weighing the values of affiliation.The experimental results show that compared with other algorithms,the method achieves better segmentation effect while maintaining good real-time performance,effectively balancing the relationship between algorithm complexity and segmentation effect.
printed materialimage segmentationsimple linear iterative clusteringfuzzy C-means clusteringsuper-pixel
万方数据
维普
