Denoising Bottleneck with Mutual Information Maximization for Video Multimodal Fusion.
Shaoxiang WuDamai DaiZiwei QinTianyu LiuBinghuai LinYunbo CaoZhifang SuiPublished in: ACL (1) (2023)
Keyphrases
- multimodal fusion
- denoising
- mutual information maximization
- high robustness
- audio visual
- video sequences
- video content
- video data
- competitive learning
- mutual information
- multimodal interfaces
- multimedia
- video frames
- relevance feedback
- image processing
- neural network
- gait recognition
- human computer interaction
- key frames
- similarity measure
- learning algorithm
- information retrieval