CAJ | 학술논문

基于区域的局部二值拟合模型在处理灰度不均匀图像方面有较大优势，但其只考虑原始图像灰度的平均统计信息，对于包含大量噪声的图像通常很难获得理想的效果。为克服上述缺陷，提出一种基于原始图像和差分图像统计信息的分割模型。该模型在原始图像灰度统计信息的基础上，加入差分图像信息，分别对原始图像和差分图像构造以高斯函数为核函数的能量方程，并运用梯度下降法求解，驱使活动轮廓向目标边缘演化。实验结果表明，与传统活动轮廓模型相比，该模型能正确提取含有噪声和信噪比低的图像，同时对初始轮廓曲线有更高的鲁棒性。
기우구역적국부이치의합모형재처리회도불균균도상방면유교대우세，단기지고필원시도상회도적평균통계신식，대우포함대량조성적도상통상흔난획득이상적효과。위극복상술결함，제출일충기우원시도상화차분도상통계신식적분할모형。해모형재원시도상회도통계신식적기출상，가입차분도상신식，분별대원시도상화차분도상구조이고사함수위핵함수적능량방정，병운용제도하강법구해，구사활동륜곽향목표변연연화。실험결과표명，여전통활동륜곽모형상비，해모형능정학제취함유조성화신조비저적도상，동시대초시륜곽곡선유경고적로봉성。
Local Binary Fitting( LBF) model based on region has great advantage in dealing with the segmentation of images with intensity inhomogeneity. Because it only considers original image statistical information,it cannot efficiently segment images with heavy noise. In order to overcome these problems,this paper proposes a novel local active contour model based on the original image and the difference image. This model that combines the difference image information is on the basis of the original image intensity statistics. The energy function is then constructed with Gaussian function as the kernel function. The contour can be driven to the objective boundaries by using the gradient descent method. Experimental results show that the proposed model is able to deal the image with noise and low signal-to-noise ratio, and can also reduce the sensitivity on the initialization of active contour when compared with the classical active contour model.