MAM: Transfer learning for fully automatic video annotation and specialized detector creation

Keyphrases

• Transfer Learning 100%
• Fully-automatic 100%
• Multiple Annotations 100%
• Automatic Video Annotation 100%
• Image Data 66%
• Object Detection 66%
• Detection Task 66%
• Eyelid 66%
• Pupil Center 66%
• Occlusion 33%
• Robot Perception 33%
• Bounding Box 33%
• Scene Understanding 33%
• Head Movement 33%
• New Dataset 33%
• Eye Tracking 33%
• Point Detection 33%
• Self-training Method 33%
• Head-mounted 33%
• Outline Detection 33%
• Head Pose Estimation 33%
• Steering Wheel 33%
• Long Head 33%
• Automatic Labeling 33%
• Car Cabin 33%
• Gaze Point 33%
• Labelled Image 33%
• Eye Pupil 33%

Computer Science

• Annotation 100%
• Transfer Learning 100%
• Object Detection 50%
• Scene Understanding 25%
• Point Detection 25%
• Multiple Method 25%
• Pose Estimation 25%