Rethinking Resolution in the Context of Efficient Video Recognition.
Chuofan MaQiushan GuoYi JiangPing LuoZehuan YuanXiaojuan QiPublished in: NeurIPS (2022)
Keyphrases
- video sequences
- recognition accuracy
- object recognition
- automatic recognition
- human activities
- activity recognition
- video database
- video data
- contextual information
- recognition rate
- computationally efficient
- action recognition
- video analysis
- higher resolution
- temporal resolution
- video content
- lower resolution
- temporal information
- neural network
- high resolution
- multiresolution
- feature extraction
- multimedia
- image processing