CAJ | 학술논문

本文从理论上分析了无需重新初始化的水平集方法的主动轮廓图像分割模型,此模型有很大的优越性,但对于目标与背景对比度较小这种情况不能得到一个好的分割效果.该模型应用于CT图像中肝脏的分割时,主动轮廓曲线会跨越肝脏边界从而导致错误的分割结果.通过修正边缘检测函数,加强了其在目标边界处的约束效果,使得主动轮廓曲线在目标物体边界处停止演化,这样能够准确的将肝脏分割出来,保证了分割的正确性.实验证明了该方法的可行性.
본문종이론상분석료무수중신초시화적수평집방법적주동륜곽도상분할모형,차모형유흔대적우월성,단대우목표여배경대비도교소저충정황불능득도일개호적분할효과.해모형응용우CT도상중간장적분할시,주동륜곽곡선회과월간장변계종이도치착오적분할결과.통과수정변연검측함수,가강료기재목표변계처적약속효과,사득주동륜곽곡선재목표물체변계처정지연화,저양능구준학적장간장분할출래,보증료분할적정학성.실험증명료해방법적가행성.
Firstly, an image segmentation method based on level set evolution without re- initialization is studied. It has relative great superiority. But we can't get a good result when the contrast of object and background is small. The active contour curve goes be-yond the liver edges in the CT image. Secondly, the method is improved by revising the edge indicator function near edges. In this case, the restriction on the edges would be strengthened,and the active contour curve will stop on the edges. Then, the edges of the object can be captured more accurately. The numerical experiments show the advantage of the improved model.