MDMMT-2: Multidomain Multimodal Transformer for Video Retrieval, One More Step Towards Generalization.
Alexander KunitsynMaksim KalashnikovMaksim DzabraevAndrei IvaniutaPublished in: CoRR (2022)
Keyphrases
- video retrieval
- video database
- video indexing
- video search
- content based retrieval
- video data
- visual content
- retrieval systems
- concept detection
- retrieval framework
- video shots
- key frames
- video content
- image and video retrieval
- semantic gap
- concept based video retrieval
- video clips
- multi modal
- video collections
- image data
- similarity measure
- interactive search
- interactive retrieval
- semantic video retrieval
- concept learning
- search engine
- information retrieval systems
- active learning
- high level
- multimedia
- computer vision