SUR-Net: predicting the satisfied user ratio curve for image compression with deep learning

Keyphrases

• Image Compression 100%
• Deep Learning 100%
• Satisfied User Ratio 100%
• Just Noticeable Difference 40%
• Fully Connected 20%
• Compression System 20%
• Direct Approach 20%
• Probability Distribution 20%
• Level Difference 20%
• Performance Prediction 20%
• Transfer Learning 20%
• Deep Learning Methods 20%
• Small Distortion 20%
• Convolutional Neural Network Features 20%
• Bhattacharyya Distance 20%
• Distortion Level 20%
• Feature Pooling 20%
• Difference Distribution 20%
• Siamese Convolutional Neural Network 20%
• Lossy Image Compression 20%

Engineering

• Just Noticeable Difference 100%
• Deep Learning Method 100%
• Good Prediction 50%
• Transfer Learning 50%
• Compression Scheme 50%
• Reference Image 50%
• Learning Approach 50%
• Prediction Performance 50%
• Direct Approach 50%
• Bhattacharyya Distance 50%
• Convolutional Neural Network 50%

Mathematics

• Deep Learning Method 100%
• Probability Distribution 50%
• Direct Approach 50%
• Lossy 50%
• Transfer Learning 50%
• Convolutional Neural Network 50%