Research And Implementation Of A Pen-based CT Image Contour Extraction System

A precise contour extraction of object organs from serial CT images provides the possibility of further medical examination, analysis, diagnose, treatment, and evaluation for the doctors. Advancing computer technology in the field of medical image analyze, especially the widely use of CT image process with computer, making computer-aid diagnose a powerful tool in medical research. GVF Snake model based contour extraction is an important research method in computer-aid CT image process.There are several necessary steps in GVF Snake based contour extraction:First, the rough contour outlined in the image by the user manually with mouse and keyboard, which then processed by specific energy function and approach to the object contour, at last reach the final result to the user confirmation. But there are problems when using mouse and keyboard as main input device:First, It's difficult to gain a rough contour close to the final result with mouse because of its lack of precise control, according to which the GVF Snake model can process with to gain a acceptable contour for doctors. Second, A text input will require the doctor to switch between the menu and main window frequently, thus losing their focus on the contour extraction or case study, lower their efficiency.By introducing a Pen-Based user interface as the main interface of the contour extraction system, we can solve the problems above through designing a Pen-Based contour extraction method, which also support a information input area to accept and recognize doctor's hand writing. Then we will approve the effectiveness and rationality of this method by experimenting a demo system to extract the contour of liver. During the software development, a Pen-Based edit tool is designed to accept the Pen-based sketch, which also support multifunction to interact with doctors, such as, pen gesture designation and recognition,text recognition. By providing a more natural interaction, this tool guarantee the doctor to focus on the contour extraction and achieved a better result.The research of this paper was supported by the Open Foundation of State Key Laboratory of Computer Science, the Chinese Academy of Sciences (No. SYSKF0704), the National High-Tech Research and Development Program of China (863 Program) (No.2006AA01Z328), the National Natural Science Foundation of China under grant (No. 60873095), and the Innovation Award of Graduate Student, Northwest University (No. 09YZZ67).