Object State Change Classification in Egocentric Videos using the Divided Space-Time Attention Mechanism.
Md Mohaiminul IslamGedas BertasiusPublished in: CoRR (2022)
Keyphrases
- space time
- attention mechanism
- video sequences
- dynamic scenes
- video representation
- input video
- spatial and temporal
- spatio temporal
- human actions
- temporal domain
- machine learning
- moving objects
- visual attention
- object tracking
- motion patterns
- image classification
- feature vectors
- feature extraction
- feature space
- multiscale
- multiple view geometry
- video objects
- high level
- video frames
- image sequences
- dynamic textures
- surveillance videos
- saliency map
- video content
- action recognition
- video data
- video synopsis