Large scale continuous visual event recognition using max-margin Hough transformation framework

Keyphrases

• Max-margin 100%
• Hough Transform 100%
• Visual Events 100%
• Event Recognition 100%
• Transformation Framework 100%
• Motion Region 50%
• Big Video 50%
• Detection Method 25%
• Extraction Methods 25%
• Video Data 25%
• Video Surveillance 25%
• Benchmark Dataset 25%
• Feature-based 25%
• Spatial-temporal Data 25%
• Local Motion 25%
• Motion Features 25%
• Preprocessing Techniques 25%
• Events of Interest 25%
• Action Recognition 25%
• Local Features 25%
• High Scalability 25%
• Region Extraction 25%
• Support Vector Machine 25%
• Hough Voting 25%
• Spatio-temporal Interest Points 25%
• Class Center 25%
• Spatio-temporal Interest Point Detectors 25%
• Motion Segmentation 25%
• Surveillance Dataset 25%
• Weighted Voting 25%
• Initial Event 25%
• Generalized Hough Transform 25%
• Activity Detection 25%
• Recognition Detection 25%
• Codebook 25%
• Region Clustering 25%
• Activity Representation 25%
• Challenging Conditions 25%
• Environmental State 25%

Computer Science

• Interest Point 100%
• Video Surveillance 50%
• Motion Feature 50%
• Action Recognition 50%
• Candidate Region 50%
• Preprocessing Step 50%
• Motion Segmentation 50%
• Hough Transforms 50%
• Support Vector Machine 50%
• local feature 50%
• Detection Method 50%
• Temporal Information 50%