摘要
心血管疾病逐步成为威胁人类生命安全的头号杀手,约有70%的致命性心血管疾病是由于心血管易损斑块破裂引起,因此,对于心血管斑块的早期检测、诊断及危险性评估具有重大的临床意义。易损斑块的准确分割是危险性定量评估的前提。由于局部容积效应的影响,传统的分割算法很难将易损斑块从周围组织中精确分割出来。MAP-EM算法通过建立组织混合模型可以很好地解决此问题,假设各组织服从高斯分布,各组织间相互独立,在最大后验概率前提下,用EM算法估计统计模型参数和组织混合比。通过和三位有经验的临床医生的十例病人手工分割结果作比较,本算法可以准确快速地将易损斑块从周围组织中分割出来。
Cardiovascular disease is the leading cause of death, and is gradually becoming the top one killer. Almost 70% of deadly cardiovascular disease is developed from the burst of vulnerable plaques. Therefore, early detection, diagnosis and risk assessment of these vulnerable plaques play an important role to prevent cardiovascular disease in the clinic. Accurate segmentation is the precondition of quantita- tive assessment. As a result of partial volume effect, it is difficult for conventional image segmentation to segment plaques from surrounding tissues. A MAP-EM algorithm that is based on tissue mixture model was investigated to provide a theoretical solution to the problem. With the assumption that each tissue type fol- lows a normal distribution and all tissue types are independent from each other, the algorithm estimated tissue mixture percentages within each image voxel and statistical model parameters for the tissue distribu- tion during each update. Ten patient datasets were used to evaluate the proposed algorithm. Comparison of the results with segmentations outlined manually by three experienced radiologists indicates an improved performance and rapid convergence of the present MAP-EM algorithm.
出处
《中国体视学与图像分析》
2009年第2期138-142,共5页
Chinese Journal of Stereology and Image Analysis
基金
国家自然科学基金(60772020)
