CAJ | 학술논문

在肺部计算机断层扫描图像的肺实质分割中，针对固定阈值法分割图像不稳定的问题，提出最佳阈值法对图像进行二值化，分割出肺实质区域，剔除肺纵膈区域及衣物等背景。与固定阈值法相比，该方法有较高的稳定性。针对由病变造成的肺部大面积坏死所引起的左右肺分开判断错误的问题补充了判断准则，直接给出结论。实验表明，该方法对肺实质分割具有较高的鲁棒性。
재폐부계산궤단층소묘도상적폐실질분할중，침대고정역치법분할도상불은정적문제，제출최가역치법대도상진행이치화，분할출폐실질구역，척제폐종격구역급의물등배경。여고정역치법상비，해방법유교고적은정성。침대유병변조성적폐부대면적배사소인기적좌우폐분개판단착오적문제보충료판단준칙，직접급출결론。실험표명，해방법대폐실질분할구유교고적로봉성。
For the Computed Tomography (CT ) scanning image segmentation of the lung parenchyma , an optimal threshold algorithm is proposed to solve the unstable problem existed in the fixed threshold method .T he optimal threshold algorithm is used to divide the image and get the lung parenchyma region after taking the lung mediastinum and clothing and other background out .The supplemental criteria are added to eliminate the misjudgment of left or right lung caused by a large area necrosis due to pathological change .Experimental results show the algorithm is more stable and effective to lung parenchyma segmentation than the others .