Variational SAR Image Segmentation Using a Local Gamma Distribution Fitting Model
Dingke Kong -, Shasha Zheng · International Conference on Electric Information and Control Engineering · 2012
Statistical distribution based geometric active contour is used as a powerful method to address SAR image segmentation problems. However, it highly depends on fitting accuracy of assumed distribution. Moreover, it has a shortage of characterizing local structure of SAR images, and cannot obtain satisfactory result of details. To overcoming the mentioned limitations, a novel active contour model based on local gamma distribution fitting energy is proposed. Experimental results of both synthetic and real SAR images verify that the model is efficient and accurate.