摘要
云计算环境下,对海量激光扫描CT图像目标特征点的快速定位,提高故障或病理诊断效率。传统的CT图像特征点定位方法采用纹理超像素边缘分割方法,当图像中激光扫描图像中出现大量的噪点和亮点分布不均衡时,图像的特征点定位准确度不高。提出一种基于活动轮廓线套索融合检索的云计算环境下海量激光扫描CT图像的快速定位算法。采用Laplace锐化算法实现激光扫描CT图像的边缘增强和锐化处理,构建图像的活动轮廓模型,对图像进行小波降噪滤波,对海量CT图像进行活动轮廓线套索融合检索的批处理,实现图像特征点快速搜索和定位。仿真结果表明,采用本文方法进行激光扫描肝脏CT图像的病斑特征点定位,实现对海量图像的批处理,在较短的迭代步数下实现准确定位,指标性能优于传统方法。
In the cloud computing environment, the rapid localization of the target feature points of the massive la-ser scanning CT image is to improve the diagnosis efficiency of fault and pathology. The traditional CT image feature point positioning method is texture segmentation method for super pixel edge. The image feature point positioning accu- racy is not high once there are massive unbalanced noisy and light points. A fast location algorithm of the massive laser scanning CT images based on cloud computing environment of active lasso retrieval is presented in this paper. The La- place sharping algorithm implemented the laser scanning CT image edge enhancement processing and sharpening, con- structed the image of the active contour model,wavelet denoised and filtered the images, batch processed the massive CT images under active lasso retrieval, realized fast reaching and location of the image feature point. Simulation result shows that the lesion feature points localization of the laser scanning liver CT image in this method realized the batch processing of massive images, achieved the accurate positioning in a short number of iterations and the performance is better htan traditional methods.
作者
蒋宏艳
闻娟
JIANG Hong-yan WEN Juan(Information Engineering Institute, Henan Radio＆TV University, Zhengzhou 450046, Chin)
出处
《激光杂志》
北大核心
2017年第4期75-79,共5页
Laser Journal
基金
2017年度河南省科技厅科技攻关基金项目(172102210126)
2015年度河南省科技厅科技攻关基金项目(152102310118)
关键词
云计算环境
激光扫描
CT图像
特征
定位
cloud computing environment
laser scanning
CT image
feature
location
