Research On Medical Image Fusion Method Based On Sparse Representation And Low-rank Decomposition

Medical image fusion is an important branch in the field of information fusion.In medicine,to achieve accurate diagnosis and help treatment of this goal.It is often necessary to study and analyze the image information from the same or different medical imaging equipment in the same lesion area under different environmental conditions.However,in the clinical diagnosis,the single image obtained is very hard to provide the doctor with comprehensive information about the pathology.The medical image fusion technique solves this problem,which provides a safe and reliable guarantee for doctors to accurately diagnose and treat diseases.Therefore,the research of medical image fusion method is a hot topic with critical theory and practical application value.Medical image fusion technology integrates the useful information in separate medical images and obtains a fusion image with richer details and more comprehensive information.Fusion technology has been successfully applied to clinical medicine at present,such as image guided radiotherapy surgery,non-invasive diagnosis and treatment plan design,has achieved significant results.Although many scholars have proposed a large number of fusion algorithms,these algorithms only apply to certain types of medical images,lack of good applicability and robustness,and which limited the development of clinical medicine to some extent.In order to solve the above problems,two medical image fusion methods are proposed.The first method using two low-rank decomposition combined with a dictionary to learn,will stay in sparse image after fusion image decomposition to decompose,effective combination of low-rank decomposition and the advantages of sparse representation for medical image fusion;The second approach is based on sparse decomposition of low rank and the significant measure of medical image fusion method,the method of integration of low rank ingredients with sparse components adopt their own strategies,to make the fused image both can retain source image of brightness information and can preserve salience characteristics as well as the details of the visual effect is good.Finally,the experimental results and analysis show that the method no matter from the Angle of human eye visual or objective evaluation index is greatly improved the medical image fusion quality and superior to the latest fusion method.