Unsupervised visual representation learning by graph-based consistent constraints

Keyphrases

• Effective Representation 100%
• Semi-supervised Learning 100%
• Positive Image 100%
• Negative Image 100%
• Consistent Constraint 100%
• Graph-based 100%
• Visual Learning 100%
• Image Pair 100%
• Unsupervised Feature Learning 100%
• Labeled Data 50%
• Visual Representation 50%
• Image Classification 50%
• Geodesic Distance 50%
• Graph Analysis 50%
• Mining Method 50%
• Competitive Performance 50%
• PASCAL VOC 50%
• Consistency Criterion 50%
• Visual Categories 50%
• Convolutional Neural Network 50%
• Classification Datasets 50%
• Unsupervised Constraints 50%
• Constraint Mining 50%
• Nearest Neighbor Matching 50%
• Positive Pairs 50%
• Supervisory Signal 50%
• Matching Pairs 50%
• Hard Negative Mining 50%
• Cycle Consistency 50%

Computer Science

• Representation Learning 100%
• Visual Representation 100%
• Semisupervised Learning 50%
• Geodesic Distance 25%
• Consistency Criterion 25%
• Image Classification 25%
• Convolutional Neural Network 25%