【摘要】 CV模型和LBF模型是两个著名的图像分割模型,然而它们有各自的缺点。CV模型不能处理灰度不均图像,而LBF模型虽然能处理灰度不均图像,但对活动轮廓的初始化很敏感,且对噪声不具有鲁棒性。为了克服上述缺点,首先对图像进行预处理,然后在得到新的图像的基础上提出与LBF类似的模型,同时将其与CV模型结合,得到全局和局部活动轮廓模型。实验结果表明,所提模型不仅能处理灰度不均匀图像,同时减弱了活动轮廓对初始化的敏感性,并且提升了对噪声的鲁棒性。更多还原

【Abstract】 CV model and LBF model are two extremely famous image segmentation models,but they all have their own shortcomings as well.The CV model cannot deal with the image with intensity inhomogeneity,while the LBF model is sensitive to the initialisation of active contours and does not robust to the noise though being able to adress the image with intensity inhomogeneity.To overcome the aforementioned faults,we first preprocess the image,and then propose a model building on the new image which is similar to LBF model.Furthermore,by integrating the CV model into the new model,we get a new active contours model based on global and local image information.The experiments show that the proposed model is able to address the image with intensity inhomogeneity,and can also reduce the sensitivity on the initialisation of active contours,and improve the robustness to noise.更多还原