Review Of Medical Image Registration

The great development of science and technology promotes artificial intelligence to combine with medical image gradually,and medical image processing has become a hot research field.Making use of advanced computer technology to process medical images has become a trend of development.Informally speaking,image registration is the operation of "aligning" two images,and only when this procedure is completed can the image be compared and analyzed.In medical image processing,image registration is the preparation procedure of many subsequent processing operations and plays a fundamental role,which means that this technology is of great significance.Through the way of literature research,this dissertation introduces the methods of image registration from three viewpoints: based on grey value,based on characteristics,based on transform-domain.Next,according to the principle of controlling other factors of registration to remain the same and keeping one factor of registration different,this dissertation designs contrastive experiments,and deals with the same data using ITK(Insight Segmentation and Registration Toolkit).Then,dissertation measures the effects of different registration methods through experimental results,and compares the advantages and disadvantages of algorithms.At last,dissertation views the existing registration methods froma new perspective,and analyzes each factor of the registration algorithm one by one.Combining with theory and experiment,dissertation classifies and summarizes the algorithms of registration,and compares advantages and disadvantages,and makes a prediction about the possible research direction of registration in the future.