基于改进深度学习的动画场景图像自动生成方法  

Automatic Generation of Animation Scene Images Based on Improved Deep Learning

作　　者：周一帆 ZHOU Yifan(School of Information Engineering,Zhumadian Vocational and Technical College,Zhumadian 463000,China)

机构地区：[1]驻马店职业技术学院信息工程学院,河南驻马店463000

出　　处：《太原学院学报(自然科学版)》2025年第2期78-84,共7页Journal of TaiYuan University:Natural Science Edition

基　　金：河南省科技攻关计划项目(212102210516)。

摘　　要：针对当前动画场景图像自动生成模式存在的精度不高、清晰度下降等问题,提出了一种基于改进深度学习的动画场景图像自动生成方法。该方法结合实际需求,通过动画场景图像虚拟重构与特征匹配处理,采用改进深度学习进行图像特征提取,并对特征进行匹配处理。设计动画场景图像自动生成模型生成动画场景图像,采用图像边缘处理的方式来优化图像质量,以此最大程度上提升图像清晰度。实验结果表明,生成的图像整体清晰度较高,更符合动画场景图像预设的自动生成标准,展示了该方法的可靠性和实际应用价值,可为动画行业的技术发展提供有效的支持和指导。To address the issues of low accuracy and decreased clarity in current automatic animation scene image generation models,a method based on improved deep learning for automatic generation of animation scene images is proposed.This method integrates practical needs by employing virtual reconstruction and feature matching processing for animation scene images.Improved deep learning techniques are used for image feature extraction,followed by feature matching.An automatic generation model for animation scene images is designed to produce these images,and edge processing techniques are applied to enhance image quality,thereby maximizing image clarity.Experimental results demonstrate that the generated images have higher overall clarity and better meet the preset standards for automatic generation of animation scene images.This showcases the reliability and practical value of the method,and thus can provide effective support and guidance for the technological development of the animation industry.

关 键 词：动画场景 图像特征 图像自动生成 改进深度学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]