An Efficient Video Prediction Recurrent Network using Focal Loss and Decomposed Tensor Train for Imbalance Dataset.
Mingshuo LiuKevin HanShiyi LuoMingze PanMousam HossainBo YuanRonald F. DeMaraYu BaiPublished in: ACM Great Lakes Symposium on VLSI (2021)
Keyphrases
- recurrent networks
- recurrent neural networks
- video data
- prediction accuracy
- video content
- video sequences
- human actions
- video frames
- weakly labeled
- multimedia
- high order
- video streams
- higher order
- video dataset
- prediction model
- feed forward
- neural network
- dimensionality reduction
- video clips
- biologically inspired
- real time
- event detection
- human activities
- video analysis
- key frames
- video retrieval
- space time
- class distribution
- diffusion tensor
- event recognition
- machine learning