非局部相似性压缩感知下的医学设备三维成像方法

Three-dimensional imaging of medical equipment based on non local similarity compressed sensing

目的针对像素间的领域结构信息难以获取,无法分析医学设备图像子块相似性的问题,本文提出了一种新的非局部相似性压缩感知下的医学设备三维成像方法。方法首先获取医学设备三维图像并对其分块处理,通过冗余字典完成稀疏图像表示,利用自回归模型模拟医学设备三维成像的非局部状态,分析三维成像的非局部相似特性,得到医学设备三维图像的局部相关性和非局部相似性,以此为限定条件,实现医学设备三维成像的相似性压缩感知。最后以某三甲医院肿瘤科的3台医疗器械为实验对象,以不同时间段产生的三维图像为实验样本,与其他两种文献方法进行了对比。结果本文提出了非局部相似性压缩感知下的医学设备三维成像方法,相比于其他两种文献方法,其压缩感知峰值信噪比与标准值的匹配度最高可达99%。结论在稳定的采样频率下获取领域结构信息,可达到最佳的图像重构效果,解决了医学设备图像子块相似性的问题,保证医学设备三维成像质量。

Objective To solve the problem that it is difficult to obtain the domain structure information between pixels and analyze the similarity of medical device image sub blocks,a new 3D imaging method of medical device based on non local similarity compressed sensing is proposed.Methods Firstly,the 3D images of medical equipment are obtained and processed in blocks.The sparse images are represented by redundant dictionaries.The non local state of 3D imaging of medical equipment is simulated by autoregressive model.The non local similarity characteristics of 3D imaging are analyzed,and the local correlation and non local similarity of 3D images of medical equipment are obtained to realize the similarity compressed sensing of 3D imaging of medical equipment.Finally,three medical devices in the Oncology Department of a 3A hospital are taken as the experimental objects,and the three-dimensional images generated in different time periods are taken as the experimental samples,which are compared with the other two literature methods.Results Compared with the other two literature methods,the peak signal-to-noise ratio of compressed sensing can match the standard value up to 99%.Conclusions Obtaining domain structure information at a stable sampling frequency can achieve the best effect of image reconstruction,solve the problem of image sub block similarity of medical equipment,and ensure the quality of 3D imaging of medical equipment.