Spatiotemporal saliency-based multi-stream networks with attention-aware LSTM for action recognition

Keyphrases

• Saliency Based 100%
• U-Net 100%
• Action Recognition 100%
• Attention-aware 100%
• Spatiotemporal Saliency 100%
• Multi-stream Network 100%
• Human Action Recognition 60%
• Video Frames 40%
• Lstm 40%
• Multi-stream 40%
• State-of-the-art Techniques 20%
• Deep Convolutional Neural Network (deep CNN) 20%
• Recognition Method 20%
• Feature Extractor 20%
• Long Short-term Memory Network 20%
• Long-range Dependence 20%
• Competitive Performance 20%
• Convolutional Neural Network Features 20%
• Optical Flow 20%
• Three-stream 20%
• Dependency Relations 20%
• Spatial Stream 20%
• UCF-101 20%
• HMDB51 20%
• Temporal Stream 20%
• Flow Frame 20%

Computer Science

• Long Short-Term Memory Network 100%
• Action Recognition 100%
• human action recognition 100%
• Experimental Result 33%
• Research Topic 33%
• Long Short-Term Memory Networks 33%
• Deep Convolutional Neural Networks 33%
• Term Dependency 33%
• Dependency Relationship 33%
• Spatial Stream 33%