Computer Vision Driven Film and TV Scene Modeling and Innovation

Abstract

This article aims to explore new methods of animation scene generation and DM (Data mining) through DL (Deep Learning) technology and provide innovative technical means and valuable data support for animation production. In order to achieve this goal, this article adopts the advanced DL model to generate animation scenes and successfully generates a large number of realistic and diverse animation scenes by training and optimizing Gan (Generative adversarial network). Moreover, this article uses DM methods such as cluster analysis and classification recognition to mine and analyze the generated animation scene data deeply. The results prove the effectiveness and feasibility of DL in animation scene generation and DM. Automatically generating animation scenes through DL and CAD (computer-aided design) data can greatly reduce the workload of manual design and production and improve the efficiency of animation production. Moreover, through DM technology, we can deeply understand the audience's preferences and market trends and provide more accurate market positioning and content innovation direction for animation production. The research results show that the animation scene generation and DM method based on DL have obvious advantages and potential.