MDMMT: Multidomain Multimodal Transformer for Video Retrieval.
Maksim DzabraevMaksim KalashnikovStepan KomkovAleksandr PetiushkoPublished in: CVPR Workshops (2021)
Keyphrases
- video retrieval
- visual content
- content based retrieval
- video database
- video data
- video indexing
- semantic gap
- video search
- multi modal
- video content
- concept detection
- video segments
- image and video retrieval
- key frames
- retrieval framework
- concept based video retrieval
- video shots
- video clips
- retrieval systems
- interactive retrieval
- semantic video
- audio visual
- image sequences
- three dimensional