The Design And Implementation Of Diagnostic Imaging System Suitable For Multi-Slice CT

Currently, multisliecs helieal CT has become one of the essential equipments in clinical imaging diagnosis. However, CT imaging diagnostic systems used for clinical are either imported three-dimensional imaging diagnostic workstation with high-end hardware configuration which are not only expensive but also incompatible with domestic PACS and other hospital information systems,or homemade workstations which have common technical problems of low diagnostic accuracy and lack of high-end computer-aided diagnostic functions such as three-dimensional reconstruction. In view of this, the topic designs and develops a new generation of CT imaging diagnostic system which is cost-effective and based on the general configuration PC. The main work includes the following three aspects:First, for the problem that a multi-slice CT examination can get thousands of CT images, while the ordinary PC configured 2G memory can not import and display thousands of two-dimensional tomographic images completely and quickly, the topic designs and implements a new image IO(Input Output) module, whose import function is characterized by parallel reading images, analysising and saving the necessary image information and classifying and sorting images based on the information, and its export function consists of a film printing and CD burning.Second, for the problem that the domestic CT imaging diagnostic systems are lack of high-end features such as 3D reconstruction, the topic designs and implements a new functional module which combines 2D and 3D images processing funtions together. Based on the existing two-dimensional image processing and image measurement functions, the module adds new two-dimensional advanced features such as the positioning line, the synchronization sequence, drawing and statistical analysis of the ROI(Region Of Interest) in any shape and so on. Besides it integrates 3D reconstruction functions such as volume rendering and surface rendering.Third, for the problem that the imported three-dimensional imaging diagnostic workstations are expensive, the topic uses OpenGL2.1 and CUDA to achieve the high speed of three-dimensional reconstruction based on a common configuration PC with GPU. It only takes about 900 ms for the volume rendering of 1700 images, and about 2s for the surface rendering of more than 650 images.