Saliency for fine-grained object recognition in domains with scarce training data

Keyphrases

• Scarce Training Data 100%
• Saliency 100%
• Fine-grained Object Recognition 100%
• Saliency Methods 80%
• Object Recognition 40%
• Saliency Map 40%
• Performance Improvement 20%
• Classification Accuracy 20%
• Image Input 20%
• Network Performance 20%
• Effective Training 20%
• Original Image 20%
• Convolutional Neural Network 20%
• Feature Extraction Methods 20%
• Limited Training Data 20%
• Attention Mechanism 20%
• Fine-grained Recognition 20%
• Recognition Model 20%
• Vision Pipeline 20%
• Convolutional Neural Network Architecture 20%
• Fine-grained Dataset 20%
• Visual Features 20%
• Performance Gain 20%

Computer Science

• Object Recognition 100%
• Training Data 100%
• Convolutional Neural Network 66%
• Large Data Set 33%
• Network Performance 33%
• Performance Gain 33%
• Visual Feature 33%
• Neural Network Architecture 33%
• Classification Accuracy 33%
• Training Sample 33%
• Extraction Process 33%
• Feature Extraction 33%

Engineering

• Object Recognition 100%
• Convolutional Neural Network 66%
• Extraction Process 33%
• Effective Training 33%
• Feature Extraction 33%
• Classification Accuracy 33%
• Recognition Model 33%
• Neural Network Architecture 33%
• Input Image 33%