Research On 3D Model Retrieval Based On Dimension Transformation

Science and technology is driving the progress of society,3D models are richer in content and attributes than in two-dimensional images,and the description of the world is more realistic.The sensory effects of two-dimensional images are gradually unable to satisfy people.So as a more lifelike simulation than two-dimensional images,3D model acquisition,rendering,propagation,We can feel the rapid development of 3D models,which is also visible to the naked eye,which makes the retrieval of 3D models become a hot topic of research.Up to now,it has been widely used in various fields.3D model has been applied in different fields.3D modeling technology and digital development technology make the process of model generation easier and easier.The main content of this paper is to study 3D model retrieval based on dimensionality transformation.Firstly,it describes the research background and significance of 3D model retrieval,and gives a brief description of the research situation at home and abroad.There are two important sub-processes in retrieval: feature extraction and its similarity calculation,in which feature extraction is the main challenge.Feature extraction is a major challenge.Describe the content of the 3D model,choose a feature to describe.Then use the combined distribution of mean curvature and thermonuclear features.After converting 3D model into 2D image based on hot kernel feature,several classical 2D image retrieval methods are introduced and compared,then SIFT algorithm and histogram feature extraction are used to match the image.Finally,the image retrieval is completed,and finally the experiment is completed.