Multivariate online kernel density estimation with Gaussian kernels
Pattern Recognition
Maximum likelihood estimation of Gaussian mixture models using stochastic search
Pattern Recognition
Hi-index | 0.00 |
Image segmentation is a fundamental step and foundation for automatic SAR image interpretation. By combining the Gabor filter bank (GFB) and active contours, this paper proposes a new SAR image segmentation method. Firstly, GFB is used to efficiently suppress speckles in SAR image and modify the gray histogram into Gaussian mixture model (GMM). Then GMM-based pixel classification is employed to pre-segment the filtered image. Finally, the active contours, initialized with the pre-segmented regions, are applied to unfiltered image for the final segmentation with several iterations. Experiments are conducted to vivificate the efficiency and effectiveness of the proposed method.