基于LoG和因子分解的集成ACM的亚实性肺结节分割

THE INTEGRATED ACM BASED ON LOG AND FACTORIZATION FOR SUB-SOLID PULMONARY NODULES SEGMENTATION

针对亚实性肺结节(SSN)边缘模糊、结构复杂、区域亮度不均等问题,传统的活动轮廓模型(ACM)方法难以达到高精度的分割结果。因此,提出一种优化的LoG算子和因子分解结合的集成ACM分割算法。用高斯拉普拉斯算子(LoG)处理肺实质图像,对算子进行优化并构建能量项,以加强边界并增强区域亮度;以像素值作为描述特征构建因子分解能量项,使曲线演变到目标边缘;将LoG能量项和因子分解能量项集成到LGIF模型中对SSN分割。实验结果证实,该算法模型对SSN的分割更有效。

Aimed at the problems of fuzzy boundary,complicated structures,and uneven regional brightness of sub-solid pulmonary nodules(SSN),the traditional active contour model(ACM)method is difficult to obtain high-precision segmentation results.Therefore,an integrated ACM model segmentation algorithm combining optimized LoG operator and factorization is proposed.The image of lung parenchyma was processed by the Laplacian of Gaussian operator(LoG),the operator was optimized and the energy term was constructed to strengthen the boundary and enhance the regional brightness.The pixels was used as a description feature to construct the factor decomposition energy term,and the energy evolved the curve to the target boundary.The LoG energy term and factorization energy term were integrated into the LGIF model to segment the SSN.Experimental results confirm that the proposed algorithm model is more effective for SSN segmentation.