欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
近年来,在医学领域里,计算机辅助诊断（CAD）技术以其高效的实用性已备受人们关注。对于X光胸片来说,肺部区域的分割是计算机辅助检测胸腔疾病的首要步骤。本研究提出一种自动的肺部分割方法。首先采用柔性形态学滤波进行初始分割,然后用结合先验知识的聚类算法进行二次分割。通过在两组不同图像（计算机X线摄影、传统X光胶片）中仿真测试,平均灵敏度和特异性可达89.39%和92.82%,该方法在保持高分割灵敏度的同时,提高了分割的特异性。
英文摘要：
Recently computer-aided diagnosis （CAD） technology has attracted more attention with its efficient practicability in medical field. Segmentation of lung fields is the principal step to detect chest disease of computeraided for chest radiographs. In this paper, an automatic segmentation method for lung fields was developed. First an initial segmentation was conducted by soft morphological filter, then a second segmentation was carried out by clustering algorithm with a priori knowledge to improve specificity and keep high sensitivity of segmentation simultaneously. Tested on two different groups of images （computed radiography, conventional X-ray film） showed that mean sensitivity and specificity were 89.39% and 92.82%. The proposed segmentation exhibited better performance and was more robust compared with other methods.
同期刊论文项目
同项目期刊论文
Robust propagation velocity estimation of gastric electrical activity by least mean p-norm blind cha
期刊信息
位置：