Research And Application Of Medical Image Registration

Image registration is to match two or more images which were taken at different time, by different sensors (imaging instrument) or at different light, climate, camera position and angle. Now image registration has been widely used in the fields of computer vision, satellite remote sensing data analysis, medical image processing which will be greater practical significance.Image registration in medicine is mainly used:a variety of medical equipment collected registration between images, which are divided into the same kind of medical registration of medical devices for the image registration that is single mode and different medical registration Equipment for registration of medical images that is multi-modal registration; image registration and image library, for future automation of the diagnosis of the disease; surgical navigation to provide information to guide the doctors wound lesions of the next step: three-dimensional space Involved in operating the navigation; observation of children's bone development and so on.Image registration algorithm consists of a rigid registration algorithm and elastic registration algorithm, in medicine, tumor size does not always change the location of the site, such as brain, bone and other tissues, which can be used with rigid registration algorithm standard treatment, commonly used rigid registration algorithm to monitor registration. However, patients with tumors in clinical cure a difficult, time changes, attached to the brain, or skeletal parts of the tumor will also change, there will be two local changes in medical images, then using elastic registration algorithm treatment, there are commonly used registration algorithm SIFT elastic, thin plate spline elastic registration algorithm. If the tumor in the abdomen and soft tissue lesions, such as:liver, lung and heart tissue, treatment, and the organization over time registration in the same manner. Elastic Registration commonly used algorithms such as SIFT elastic registration plate spline algorithm and elastic registration algorithms can deal well deformed lesions in the same local area image, but processing time is longer, does not have a local variable nature of The B-spline has the property, while Powell optimization algorithm without calculating the time in the iterative gradient, can accelerate the searching speed, which makes the Powell optimization algorithm based on B-spline elastic registration has to speed up the rate of registration, experimental results show that the algorithm can deal with medical images.Image registration has been used for image processing equipment, such as:the traditional radiation therapy equipment, has the function of the CT image-guided machines and map with image guided function of X-ray machine. But these devices have a corresponding disadvantages are:the traditional radiation equipment with the patient radiation for a long time, lesion localization and other shortcomings are not allowed; CT features with image-guided machines and map with image guided function of X-ray machine with a single feature of the device will make slice image texture is not clear or do not see the physical characteristics, the doctor can not accurately according to the image positioning information and precise removal of lesions lesions. Then made to improve these shortcomings image guided medical robots, the robot is a function of image guided medical devices sublimation, but also low-cost surgery is minimally invasive surgery and the scope of early detection of tumor lesions using image guided medical robot therapy, able to accurately locate the site of lesions to help doctors develop operation plans to guide doctors in the surgery operation, making the surgery caused the sores on the smallest of patients, secondary damage to a minimum. The accuracy of lesion localization of tumor radiation therapy is the key. the core of the image registration is achieved, in the second section of this chapter gives the image registration in medical robots processes.