Sports Video Captioning by Attentive Motion Representation based Hierarchical Recurrent Neural Networks.
Mengshi QiYunhong WangAnnan LiJiebo LuoPublished in: MMSports@MM (2018)
Keyphrases
- recurrent neural networks
- sports video
- temporal segmentation
- feed forward
- echo state networks
- neural network
- image sequences
- visual cues
- motion estimation
- event detection
- genetic algorithm
- human motion
- image representation
- artificial neural networks
- video analysis
- mid level
- visual attention
- visual content
- video content
- human body
- video clips
- optical flow
- moving objects
- knowledge base