Cluster Ensemble Based Image Segmentation Algorithm
Lei Wang, Guoyin Zhang · 2015
By integrating image segmentations to form a final image cluster is an effective image process, called image cluster ensemble. The image processing improves the accuracy and stability from traditional single clustering based algorithm. In the paper, we design a novel image partition algorithm, called Cluster Ensemble algorithm by using the K-Means and Nyström Spectral Clustering (CEKMNSC). The algorithm requires low computational complexity. It adopts a cluster ensemble scheme. In clustering, the algorithm uses a k-means algorithm to create a set of segmentations results. In ensemble processing, the algorithm integrates partition results based on the Nyström method. Our experimental results show that CEKMNSC algorithm has higher quality of clustering.