一种基于门机制多模态融合的情感分析方法

• Format: Patent
• Language: Chinese
• Published: 18.04.2023
• Subjects: ACOUSTICS; CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

本发明涉及到多模态情感分析领域，提出了一种基于门机制多模态融合的情感分析方法，具体为：首先获取多模态情感分析数据，针对不同模态特点提取特征表示；通过不同模态特征的表示获得每个模态特征对应的模态信息门；使用不同模态信息门对各模态进行模态增强，丰富模态信息表示；拼接不同模态表示得到多模态融合表示，使用多模态融合表示进行情感分类。本发明可以更好地增强主体情感表达模态的作用，削弱冗余情感表达模态的负面影响，帮助多模态特征情感识别更加精确。 The invention relates to the field of multi-modal sentiment analysis, and provides a sentiment analysis method based on gate mechanism multi-modal fusion, and the method specifically comprises the steps: firstly obtaining multi-modal sentiment analysis data, and extracting feature representation for different modal features; obtaining a modal information gate corresponding to each modal feature through representation of different modal features; carrying out modal enhancement on each modal by using different modal information gates to enrich modal information representation; splicing differentmodal representations to obtain multi-modal fusion representations, and using the multi-modal fusion representations to perform sentiment classification. According to the method, the effect of the subject e