期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2017
卷号：6
期号：7
页码：13318
DOI：10.15680/IJIRSET.2017.0607091
出版社：S&S Publications
摘要：segmentation is a core technique in medical image analysis. By enabling semantic interpretabilityof raw image data, it allows both medical experts and automatic processing software to diagnose and to plan for thetreatment of pathologies with greater accuracy and efficiency. The level set method (LSM) can be a capable techniquehowever rapid process employing correct segments is still complex task. The region based model does inadequately forintensity irregularity images. For this, we have a whole new tendency to propose a better region based level set methodof which integrates the altered signed pressure function because of the geodesic active contour models plus theMumford-Shah model. Therefore as to eliminate the re-initialization method for ancient level set model and removesthe computationally costly re-initialization. A compared employing ancient model, our model is more stable againstimages employing weak edge and also intensity irregularity. The novelty contained by our method is to help youlocally compute improved Signed pressure power (SPF) function, which uses neighborhood mean values which allow itto detect boundaries within the homogenous places. Compared with other active design models proposed methodderives valuable advantages not having difficulties just using quick process, automation and also correct medical imagesegments. This method offers undergone several analysis tests to prove its importance in medical image segmentation.
关键词：Active contour models; Chan Vese model; Geodesic active contours; Intensity irregularity; Level set;method; Medical image segmentation; Mumford-shah model; Modified Signed pressure force; Region Based;Segmentation
