Video BagNet: short temporal receptive fields increase robustness in long-term action recognition.
Ombretta StrafforelloXin LiuKlamer SchutteJan van GemertPublished in: ICCV (Workshops) (2023)
Keyphrases
- action recognition
- receptive fields
- human actions
- space time
- spatial temporal
- action classification
- motion history images
- spatial and temporal
- spatio temporal
- video dataset
- action detection
- biologically inspired
- recognition of human actions
- natural images
- recognizing human actions
- video sequences
- atomic actions
- activity recognition
- image representation
- temporal information
- human activities
- computer vision
- space time interest points
- video frames
- temporal structure
- temporal resolution
- hidden layer
- artificial neural networks
- multiscale
- feature selection
- video data
- image quality
- temporal relations
- pairwise
- three dimensional
- multimedia
- image processing
- neural network