Research on Simulation and Reconstruction of Digital Sculpture 3D Models Based on Deep Learning Algorithms

Film and television production began using digital carving software, which sculpted character images directly on a computer using a mouse or pressure pen. In the traditional production process of industrialization of film and television works, concept designers are usually only responsible for one or two preliminary tasks such as hand drawn sketches, which makes it impossible for designers to directly intervene and can only be achieved by visual effects implementation teams. This article attempts to simulate and reconstruct a digital sculpture 3D (Three Dimensional) model based on the DL (Deep Learning) algorithm. Finally, through experimental verification and analysis, it is shown that the average recognition rate of the algorithm proposed in this paper can reach 91.36%, and its performance is superior to the other two algorithms. However, the recognition rate of ordinary SAR image source algorithms is only 66.68%. The recognition rate of the enhanced Lee filtering algorithm is 82.14%. However, for the simulation time, as the recognition algorithm presented in this article adopts a new recognition algorithm based on a hybrid network of multi-layer RBM and GRNN, its simulation time is relatively long. This method can effectively improve the correct recognition rate of targets and is an effective SAR image target recognition method. Implement abstract description and classification of input images, and simulate and reconstruct real digital sculpture 3D models.