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：2018
卷号：7
期号：1
页码：489
DOI：10.15680/IJIRSET.2018.0701082
出版社：S&S Publications
摘要：Image segmentation is a pre-requisite to medical image analysis. Medical image segmentation has longbeen recognised as a difficult problem. In Medical images, accurate segmentation is a key step in contouring duringradiotherapy planning. Magnetic Resonance Imaging (MRI) are the most widely used radiographic techniques indiagnosis, clinical studies and treatment planning. In this paper, the technique incorporated is a Reverse AccuracyClassification (RCA) framework that predicts the performance of a segmentation method on new data. RCA takes thepredicted segmentation from a new image to train a reverse classifier that is evaluated to a set of reference images withavailable Ground Truth. The ideal concept is that to rank the best segmentation results in the database without makingthe manual label. Then match the rank between the predictions of the truth saved in database. Rather than going in atraditional way of classification, a reverse way of testing is followed which lead to the prediction of segmentationquality in the absence of Ground Truth.
