基于mini_Xception的动漫人物表情识别  

Anime Character Expression Recognition based on mini_Xception

作　　者：梁博文 LIANG Bowen(Xinjiang University,Urumqi Xinjiang 830000,China)

机构地区：[1]新疆大学,新疆乌鲁木齐830000

出　　处：《信息与电脑》2023年第15期158-161,共4页Information & Computer

摘　　要：动漫作为一种常见的艺术形式,与现实生活有着密切的关系。艺术家通常会从现实生活中汲取灵感,并将其转化为自己的创作素材。基于传统人脸表情分类方法,首先收集并制作包含开心、生气、无表情3种情绪的数据集,其次基于mini_Xception网络训练表情识别模型并通过大量实验验证其有效性,最后总结了动漫表情识别研究存在的问题和未来发展方向,以期为动漫人物情感分析提供借鉴。Animation,as a common art form,has a close relationship with real life.Artists often take inspiration from real life and turn it into their own material.Based on the traditional facial expression classification method,this paper firstly collects and produces three kinds of emotion data set,including happy,angry and blank expression.Secondly,it trains the expression recognition model based on mini_Xception network and verifies its effectiveness through a large number of experiments.Finally,it summarizes the existing problems and future development direction of animation expression recognition research.In order to provide reference for the emotion analysis of cartoon characters.

关 键 词：深度学习 mini_Xception 面部表情识别 动漫表情数据集 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]