Improvement and Application of Fusion Scheme in Automatic Medical Image Analysis

The research in this paper provides generalization and new ideas for topics computer-assisted medicine. main improvement efforts deep learning-based multimodal fusion schemes, which provide alternative directions robust feature fitting performance are building complex structures, migrating knowledge or experience, processing enhancing data, targeting features semantic correction based on contextual features. At the application level, brain, liver, lungs targets of scientific research, so surveys related work analyzes reasons gains. Taken together, image schemes can assist physicians understanding information about lesion sites, types, sizes, providing an important basis developing personalized treatment plans, is improving diagnosis specifying precise plans. Therefore, investigation medical promising beneficial.