Design And Implementation Of The Medical Image Processing And Analysis Software Platform

With the rapid development of medical imaging equipment, image processing algorithms also emerge in an endless stream for different imaging equipment. In order to make full use of the existing algorithms, to avoid duplication of development, some domestic and foreign research institutions have developed medical image processing and analysis software platform. Existing medical image processing and analysis software platform in some particular fields has achieved great success, but there are still problems as follows: some powerful medical image processing and comprehensive analysis platform require supporting medical equipment or special hardware to run, they are too expensive to learning and commonly using in medical field; open source framework platform software secondary development learning cost is too high, this mainly due to they rely on the open source software libraries are too large and complex; Because of the complexity of the medical image data itself, the existing software platforms are too professional, lack of generality, and inconvenient to use.The design goal of this thesis is to integrate common medical image processing and analysis algorithms in Windows operating system, reduce the learning cost in tools and algorithms use for algorithm researchers, making them to focus on the design and comparative analysis of the new algorithm. The main contents of this thesis are as follows:Firstly, design of the medical image processing and analysis algorithm platform. The algorithm platform is divided into three layers: the first layer, bottom layer, based on VTK, ITK, FSL and some mature open source algorithms; the second layer, middle layer, accordance with the unified medical file read and write interface to encapsulation bottom layer algorithms which about medical image format conversion, filtering, registration, segmentation and 3D reconstruction algorithms, easy to invoked by application layer; the third layer, application layer, the middle layer provides the interface algorithm according to certain rules of combination and connection, design the image format conversion, registration and segmentation, display three interaction module.Secondly, implementation of the medical image processing and analysis software. According to the design of medical image processing and analysis algorithm platform, encapsulate the correlation algorithm in the bottom layer of ITK and VTK algorithm library, FSL algorithms library on image registration and segmentation of brain structural modules have been transplanted to run under Windows operating system. The MATLAB environment running hippocampus 3D segmentation algorithm have been integrated into this algorithm platform, all algorithms in accordance with the rules of the middle layer design of the medical file read write interface, software has achieved three interactive module of application layer: first, image format conversion module, conversion of DICOM slice series in 3D NIf TI file, conversion 3D NIf TI file in DICOM slice series and others image format conversion functions; second, registration and segmentation module, achieved the FLIRT(FMRIB's Linear Image Registration Tool), FIRST(FMRIB's Interesting Region Segment Tool), BET(Brain Extraction Tool), HST(Hippocampus Segment Tool) four algorithms; third, the display module, achieved the 2D medical image display and 3D reconstruction, 2D display support ROI sketch, 3D reconstruction support surface, mesh, volume rendering, the volume rendering support label image overlay display.Thirdly, testing of the software functions. Using MRI and CT images to test the format conversion module, registration and segmentation module, display module, verified the correctness of the three modules. In particular, using 18 groups of brain MRI images from ADNI database to test FIRST algorithm under the Windows and Linux operating systems, according to the Precision similar measure to estimate the results of FIRST algorithm.The segmentation results has no obvious difference between two operating systems, this proved that the FIRST algorithm execute correctly after transplanted from Linux operating system.