Action recognition using multilevel features and latent structural SVM

Computer Science

• Action Recognition 100%
• Support Vector Machine 100%
• Gaussian Mixture Model 100%
• Points of Interest 66%
• Classifier 33%
• High Correlation 33%
• Recognition Accuracy 33%
• Youtube 33%
• Video Training 33%
• Scale-Invariant Feature Transform 33%
• Motion Pattern 33%
• Level Visual Feature 33%
• human action recognition 33%
• Temporal Context 33%
• Temporal Relationship 33%
• Background Model 33%
• Interest Point 33%