摘要
艺术图像的视觉情感分析有助于对艺术图像的鉴赏与保护。本文利用改进的FPN模型提取艺术图像的不同层次的情感语义特征,运用CLAHE突出艺术图像的颜色特征,分别提取图像在HSV颜色空间下的H-S二维特征及在YCrCb颜色空间下的CrCb二维特征,并设置两种颜色特征的不同权重,提出一种基于图像特征融合的情感识别模型HCFNet。实验结果表明,本文提出的模型对艺术图像情感识别的准确率可达到90.63%,比经典卷积神经网络模型和未改进FPN模型均有提升,可有效实现艺术图像情感识别任务。
The visual emotion analysis of artistic images is helpful to the appreciation and protection of artistic images.The improved FPN model is used to extract different levels of emotional semantic features of art images,and CLAHE is used to highlight the color features of art images.The H-S two-dimensional features of images in HSV color space and the crcb two-dimensional features in YCrCb color space are extracted respectively,and the different weights of the two-color features are set,an emotion recognition model hcfnet based on image fusion is proposed.The experimental results show that the accuracy of the proposed model for art image emotion recognition can reach 90.63%,which is improved compared with the classical convolutional neural network model and the unmodified FPN model,and can effectively complete the task of art image emotion recognition.
作者
杨松
刘佳欣
潘建达
YANG Song;LIU Jiaxin;PAN Jianda(School of Software,Dalian University of Foreign Languages,Dalian Liaoning 116044,China;Language Intelligence Center,Dalian University of Foreign Languages,Dalian Liaoning 116044,China;Research Center for Networks Space Multi-Languages Big Data Intelligence Analysis,Dalian Liaoning 116044,China)
出处
《智能计算机与应用》
2022年第1期146-154,共9页
Intelligent Computer and Applications
基金
国家自然科学基金(61806038)
辽宁省社会科学规划基金(L18BTQ005)
辽宁省教育厅科学研究项目(2019JYT07)
关键词
艺术图像
视觉情感分析
颜色特征
特征融合
artistic image
visual sentiment analysis
color features
features fusion
