Improving Compressed Video Using Single Lightweight Model with Temporal Fusion Module †

Keyphrases

• Lightweight Model 100%
• Compressed Video 100%
• Fusion Module 100%
• Temporal Fusion 100%
• Video Quality 66%
• Number of Parameters 66%
• Color Channel 66%
• Neural Network Model 33%
• Video Frames 33%
• Original Video 33%
• Degree of Compression 33%
• Quantization Parameter 33%
• Learning-based 33%
• Hardware Devices 33%
• Single Model 33%
• BD-PSNR 33%
• Temporal Correlation 33%
• Number of Bits 33%
• BD-BR 33%
• High Compression Ratio 33%
• Chromaticity 33%
• Receptive Field 33%
• Restoration Method 33%
• Dilated Convolution 33%
• Recursive Neural Network 33%
• Compression Artifacts 33%
• Video Compression Algorithms 33%
• Visual Artifacts 33%
• HEVC Test Model 33%

Engineering

• Color Channel 100%
• Network Model 50%
• Recursive 50%
• Objective Function 50%
• Compression Algorithm 50%
• Model Test 50%
• Temporal Correlation 50%
• Compression Ratio 50%
• Receptive Field 50%

Computer Science

• Recursive Neural Network 25%