Parallel multi-head dot product attention for video summarization.
Bohdan BilonohSergii MashtalirPublished in: DSMP (2020)
Keyphrases
- video summarization
- dot product
- audio visual
- video content
- video summaries
- video data
- event detection
- similarity function
- key frames
- real time
- visual attention
- video retrieval
- feature space
- gaussian kernels
- kernel function
- surveillance videos
- object recognition
- video sequences
- data sets
- video frames
- video streams
- multi modal
- image classification
- low level features