一种基于改进的遗传算法的脑MR图像去偏移场模型

Brain MR images de-bias model based on adapted genetics algorithm model

由于磁共振图像(Magnetic Resonance Images,MRI)常含有偏移场,影响后继图像分割。采用Legendre多项式基函数来拟合偏移场,以去除偏移场对图像分割的影响。当使得恢复图像的信息熵达到最小时,求得的偏移场最优。求偏移场的过程中需要求解基函数的参数,由于传统的梯度下降法易陷入局部最优,将遗传算法引入到参数求解过程中,然而传统的遗传算法时间复杂度高,易陷入局部最优,对遗传算法进行了改进,更容易得到全局最优解且时间复杂度较低。实验证明该算法可以得到精确的偏移场,得到准确的分割结果。

Intrascan intensity in homogeneities is a common source of difficulty for MRI segmentation.The authors estimate the bias field by Legendre polynomials to find the parameters with minimum entropy,conventional ways such as gradient-descent method often find local best,to find global best,the authors present genetics algorithm to find best parameters to estimate the bias field,but it can not always find global best neither.Then the authors make some modification of genetic algorithm to make it easier to find global best.Experiments on the segmentation of brain magnetic resonance images show the modification in this paper can find more accurate bias field and have better effect in image segmentation.