| Along with the trend of digital hospital construction, digital imaging technology is bound to develop rapidly, the traditional film management methods of the medical imaging will be eliminated inevitably, Although this can be able to reduce a lot of human resources, material resources and financial resources and so on, the facing problem must to be solved are the storage, transmission, display and management of numerous different format medical images from different medical imaging equipment. PACS (Picture Archiving and Communication System) is an important part of hospital information system, the main purpose of PACS is proposed to solve these issues. However, since the1970s, because the development of digital imaging technology, a variety of imaging equipment continue to emerge, which caused many manufacturers to independent research medical imaging device, and different manufacturer’s devices have different image storage format and transmission mode, and different devices could not be interconnected, so the resource of image information could not be shared in different information systems and departments. DICOM (Digital Imaging and Communications in Medicine) was designed as the standard for medical digital images’storage and transmission, which has become the international standard, and now it is complied with by the provider of vast majority medical imaging equipment and information systems of the world, it’s the foundation of the promotion and implementation of PACS system.In the paper, lung nodule detection in CT image of DICOM format is researched as the practical application background, an improved method of template matching is proposed to against the conventional Normalized Cross-Correlation algorithm, which is not satisfied with the time of matching. The new method improves the real-time performance of the algorithm by optimize the search strategy, which is not only ensure the accuracy of the template matching, but also greatly reduce the matching time, effectively improve the efficiency of lung nodule detection. Although, DICOM is one of the most commonly used format of medical images, it is also a special kind of image format, it can’t be displayed in the Windows platform directly, which can be recognized by the special software. Currently, a large number of medical equipment are bought on abroad in many hospital, the image processing software are mostly installed on a UNIX-based graphics workstation, if the software is purchased separately, the price is very expensive, even equivalent to the price of the equipment, and many restrictions and inconveniences are not suitable for the small and medium-sized hospital in our country. For these reasons, research and develop the medical image processing software independently based on the Windows platform is very important and meaningful, it is not only conducive to the development of remote medical diagnosis, but also beneficial to teaching and research.This paper mainly consists of two parts, the first part is a detailed analysis of DICOM format is given by intensive study of DICOM3.0standard, mainly according to the object-oriented features, the format transformation from DICOM image to BMP image is realized by using the C++programming in the visual studio2005, it includes image display, dynamically adjust window-leveling and window-width, conventional geometric transformation and so on. In second part, an improved template matching method is proposed after the segmentation of lung parenchyma image. The new algorithm is in accordance with the coarse-fine match thinking, it finds all possible matching points firstly, then find the best matching points in the neighborhood of every possible matching point, the research laid a foundation for medical image matching based on DICOM3.0standard. |
PDF Full Text Request