ReELFA: A scene text recognizer with encoded location and focused attention

Keyphrases

• Text-to-image 100%
• Focus of Attention 100%
• Scene Text 100%
• Location Attention 100%
• Recognizer 100%
• Benchmark Dataset 33%
• Spatial Data 33%
• Attention Mechanism 33%
• Spatial Relationship 33%
• Feature Map 33%
• Pooling Operation 33%
• One-hot 33%
• Scene Text Detection 33%
• Feature Area 33%
• Focus Drift 33%
• Drift Problem 33%
• Noisy Text 33%
• Curved Text 33%
• Feature Regions 33%
• 2D Features 33%
• Recognition Performance 33%

Computer Science

• Recognizer 100%
• Focused Attention 100%
• Long Short-Term Memory Network 66%
• Attention (Machine Learning) 33%
• Spatial Information 33%
• Spatial Relationship 33%
• Recognition Performance 33%
• Region Feature 33%
• Benchmarking Dataset 33%
• Focus Attention 33%

Engineering

• Recognizer 100%
• Spatial Information 33%
• Conducted Experiment 33%