Adaptive Ensemble Clustering for Image Segmentation in Remote Sensing

Image segmentation is a fundamental computer vision task. Although many approaches have been proposed, obtaining accurate results in some special applications are still not easy. In this paper, we propose a novel image segmentation method for remote sensing based on adaptive cluster ensemble learning. The clustering parameter of each image is calculated with affinity propagation automatically. Then, multiple clusterers are trained separately and the predictions of them are combined under the ensemble learning framework. In this way, the robustness of each clusterer could be enhanced. Experimental results demonstrate the effectiveness of our proposed method.