CAJ | 학술논문

目的：本文通过多种图像分割技术将损伤区域准确地标记出来,并计算出法医图像的损伤区域面积。方法：利用图像分割方法分割目标和背景,然后用数学形态学处理来消除分割带来的空洞和毛刺,最后通过比较4种基于区域分割的方法,给出了计算法医损伤面积的最佳方法。结果：迭代阈值、最大类间方差法和K-均值聚类3种分割方法都能较好地实现损伤图像的分割,同时得到误差小于0.50%的损伤面积。结论：在MATLAB平台上,利用基于区域分割的多种图像方法可以有效地计算出不规则法医损伤区域的面积。
목적：본문통과다충도상분할기술장손상구역준학지표기출래,병계산출법의도상적손상구역면적。방법：이용도상분할방법분할목표화배경,연후용수학형태학처리래소제분할대래적공동화모자,최후통과비교4충기우구역분할적방법,급출료계산법의손상면적적최가방법。결과：질대역치、최대류간방차법화K-균치취류3충분할방법도능교호지실현손상도상적분할,동시득도오차소우0.50%적손상면적。결론：재MATLAB평태상,이용기우구역분할적다충도상방법가이유효지계산출불규칙법의손상구역적면적。
Objective：This paper develops a method to tag the injured region and calculate the area of this region in medical jurisprudence image based on many species of image segmentation methods.Methods： Segmentation methods are used to partition the foreground and background and then the mathematical morphology processing is used to eliminate the holes and spots on the image.At last this paper proposes the best methods to calculate the area of the injured region by comparison of iteration threshold segmentation,Otsu method,Local threshold segmentation,K-means clustering segmentation.Results： Iteration threshold segmentation,Otsu method,and K-means clustering segmentation are effective to segment the medical jurisprudence image and the errors of the injured area are less than 0.50%.Conclusion： The region-based image segmentation algorithms running on MATLAB platform can effectively calculate the area of irregular region in the medical jurisprudence image.