Real-time head pose estimation by tracking and detection of keypoints and facial landmarks

Keyphrases

• Head Pose 100%
• Head Pose Estimation 100%
• Facial Landmarks 100%
• Kalman Filter 40%
• Detection Scheme 40%
• Tracking Schemes 40%
• Head Rotation 40%
• Long Head 20%
• State Transition Matrix 20%
• Covariance 20%
• Varying Illumination 20%
• Prediction Steps 20%
• 3D Projection 20%
• Correction Step 20%
• Simple Geometric Model 20%
• Real-time Capability 20%
• Fusion Framework 20%
• Triangulation 20%
• Novel Fusion 20%
• 2D-3D Correspondence 20%
• Monocular Video 20%
• Landmark-based 20%
• Video Data 20%
• Intensity Image 20%
• Keypoint-based 20%
• Publicly Available Datasets 20%
• Reprojection Error 20%
• Partially Occluded 20%
• Mean Average Error 20%
• 2D Keypoints 20%

Computer Science

• Pose Estimation 100%
• Kalman Filter 100%
• Varying Illumination 50%
• Detected Landmark 50%
• State Transition 50%
• Consecutive Frame 50%
• Image Intensity 50%
• Geometric Model 50%
• Transition Matrix 50%

Mathematics

• Kalman Filtering 100%
• Transition Matrix 50%
• Covariance 50%