Self-taught learning of a deep invariant representation for visual tracking via temporal slowness principle

Keyphrases

• Appearance Changes 25%
• Computer Vision 25%
• Discriminative Tracking 25%
• Image Patch 25%
• Invariant Representation 100%
• Learning Representations 25%
• Local Representation 25%
• Logistic Regression Classifier 25%
• Method Performance 25%
• Motion Pattern 25%
• Negative Samples 25%
• Observation Model 75%
• Online Training 25%
• Particle Filter Framework 25%
• Self-taught Learning 100%
• Sequential Images 25%
• Slowness Principle 100%
• Specific Information 25%
• Stacked Convolutional Autoencoder 25%
• Target Motion 25%
• Tracker 75%
• Tracking Method 25%
• Training Samples 50%
• Unlabeled Data 25%
• Visual Representation 50%
• Visual Tracking 100%

Computer Science

• Autoencoder 50%
• Computer Vision 50%
• Experimental Result 50%
• Invariant Representation 100%
• Logistic Regression 50%
• Motion Pattern 50%
• Particle Filter 50%
• Tracking Method 50%
• Training Sample 100%
• Unlabeled Data 50%
• Visual Representation 100%