计算机技术与发展2024，Vol.34Issue(10)：171-177,7.DOI:10.20165/j.cnki.ISSN1673-629X.2024.0209
基于目标对齐和语义过滤的多模态情感分析
Multimodal Sentiment Analysis Based on Target Alignment and Semantic Filtering
摘要
Abstract
In recent years,many studies have utilized attention mechanisms to capture visual representations corresponding to opinion targets for sentiment prediction,but such methods are not ideal for fine-grained opinion target alignment.To address this,a multimodal sentiment analysis method based on target alignment and semantic filtering is proposed.First,the target recognition method Deepface is introduced to obtain coarse-grained opinion targets from images,and a mapping method is used to map these coarse-grained opinion targets to fine-grained opinion targets,achieving intra-modal target alignment.Second,emotion words associated with coarse-grained opinion targets obtained by Deepface are fused with visual representations,enabling the model to more accurately understand and represent the emotional tendencies of opinion targets.Finally,the text-image matching model CLIP is introduced to evaluate the semantic correlation between images and opinion targets,thereby filtering out redundant visual modal data noise.Experiments demonstrate that the proposed opinion target alignment and semantic filtering can better utilize visual modal information and improve the accuracy of sentiment prediction.关键词
方面级情感分析/目标对齐/语义过滤/噪声/多模态Key words
aspect-based sentiment analysis/target alignment/semantic filtering/noise/multimodal分类
信息技术与安全科学引用本文复制引用
欧阳梦妮,樊小超,帕力旦·吐尔逊..基于目标对齐和语义过滤的多模态情感分析[J].计算机技术与发展,2024,34(10):171-177,7.基金项目
新疆维吾尔自治区自然科学基金项目(2022D01A99) （2022D01A99）
国家自然科学基金项目(62066044,62167008) （62066044,62167008）
新疆师范大学2022 年度青年拔尖人才项目(XJNUQB2022-23) （XJNUQB2022-23）
