Skeleton-Based Action Recognition With Low-Level Features of Adaptive Graph Convolutional Networks.
Jialin GangYao XiaoShenglan LiuYao LuPublished in: IEEE Access (2021)
Keyphrases
- action recognition
- low level features
- mid level
- mid level features
- low level descriptors
- high level
- low level
- bag of words
- human actions
- semantic gap
- visual features
- semantic information
- high level semantics
- activity recognition
- computer vision
- higher level
- high level features
- image representation
- human detection
- relevance feedback
- graph structure
- image retrieval
- key frames
- semantic attributes
- image annotation
- bag of visual words
- visual words
- visual information
- semantic concepts
- semantic content
- information retrieval
- multiscale
- atomic actions
- learning algorithm