SAR image segmentation using quantum clonal selection clustering

A novel clustering algorithm is proposed, which is derived from physical intuition of quantum mechanics and biological principle based on immune clonal selection. As extension ideas of scale-space clustering and support vector clustering, quantum clustering method deduces the clustering allocation by gradient descent, which is prone to getting stuck in local extremes. By designing a novel and high-efficiency affinity function, we adopt an immune clonal selection algorithm with elite preservation strategy to search the global optimum. The experimental results on texture images and SAR images segmentation we demonstrate show that quantum clonal selection clustering method performs well both in precision and efficiency.