基于改进测地线轮廓模型的图像分割算法

针对测地线轮廓模型对初始轮廓的敏感度高的情况,提出一种改进的测地线轮廓模型。该模型在测地线轮廓模型的基础上,引入区域方差概念,加快了初始轮廓的演化效果;同时,在几何化模型的能量函数中加入判别函数,进一步提高了分割速度与目标背景区分度。仿真实验结果表明,改进测地轮廓模型较原模型对初始轮廓的放置有很好的鲁棒性,是一种有效的图像分割算法。

Image segmentation algorithm based on an improved geodesic contour model

Aiming at the geodesic contour model to the sensitivity of the initial contour is high, this paper puts forward an improved geodesic contour model, this model using regional variance concept, speeding up the initial contour evolution effect is based on the geodesic contour model, at the same time in geometrization model of energy function, join discriminant function, further improve the speed of segmentation and target differentiate background. The simulation results show that the improved geodesic contour model has a good robustness, and it is better than the original model in the placement of initial contour.