Counterfactual Cross-modality Reasoning for Weakly Supervised Video Moment Localization.
Zezhong LvBing SuJi-Rong WenPublished in: ACM Multimedia (2023)
Keyphrases
- weakly supervised
- object localization
- weakly labeled
- video data
- topic models
- relation extraction
- superpixels
- video sequences
- object class
- named entities
- semi supervised
- video frames
- knowledge base
- image processing
- object detectors
- key frames
- computer vision
- semantic relations
- learning algorithm
- question answering
- information extraction
- natural images
- image classification
- denoising