SALICON: Reducing the semantic gap in saliency prediction by adapting deep neural networks

Keyphrases

• Saliency 100%
• Deep Neural Network 100%
• Semantic Gap 100%
• Saliency Prediction 100%
• Saliency Model 66%
• Human Behavior 33%
• Benchmark Dataset 33%
• Model Prediction 33%
• Information Integration 33%
• Eye Tracking 33%
• Function-based 33%
• Architecture-centric 33%
• Object Recognition 33%
• Evaluation Metrics 33%
• Visual Attention 33%
• Large Gap 33%
• High-level Semantics 33%
• Image Statistics 33%
• Challenge Set 33%
• Image Level 33%
• Strong Semantics 33%
• Metric Information 33%
• Semantic Content 33%
• Representational Power 33%
• Eye Fixation 33%
• Image Scale 33%
• Human Fixation 33%
• Saliency Ratio 33%
• Tracking Benchmark 33%

Computer Science

• Deep Neural Network 100%
• semantic gap 100%
• Object Recognition 20%
• Evaluation Metric 20%
• Higher Semantic Level 20%
• Stronger Semantics 20%
• Objective Function 20%
• Model Prediction 20%
• Visual Attention 20%