Sign in

CorMulT: A Semi-supervised Modality Correlation-aware Multimodal Transformer for Sentiment Analysis

By Yangmin Li and others
Multimodal sentiment analysis is an active research area that combines multiple data modalities, e.g., text, image and audio, to analyze human emotions and benefits a variety of applications. Existing multimodal sentiment analysis methods can be classified as modality interaction-based methods, modality transformation-based methods and modality similarity-based methods. However, most of... Show more
July 9, 2024
=
0
Loading PDF…
Loading full text...
Similar articles
Loading recommendations...
=
0
x1
CorMulT: A Semi-supervised Modality Correlation-aware Multimodal Transformer for Sentiment Analysis
Click on play to start listening