HaltingVT: Adaptive Token Halting Transformer for Efficient Video Recognition.
Qian WuRuoxuan CuiYuke LiHaoqi ZhuPublished in: CoRR (2024)
Keyphrases
- object recognition
- human activities
- recognition accuracy
- recognition rate
- computationally efficient
- video data
- video frames
- visual recognition
- video database
- video analysis
- scalable video
- multimedia data
- feature extraction
- online video
- real time
- real time video
- power transformers
- video content
- event detection
- multimedia
- neural network