ViNet: Pushing the limits of Visual Modality for Audio-Visual Saliency Prediction

Computer Science

• Visual Feature 100%
• Human Performance 100%
• Action Recognition 100%
• Deep Learning Method 100%
• Trained Network 100%
• Visual Modality 100%
• Prediction Model 100%
• Pre-Trained Model 100%
• Trilinear Interpolation 100%

Keyphrases

• Visual Modality 100%
• Audio-visual Saliency 100%
• Saliency Prediction 100%
• Audio 60%
• Decoder 40%
• Audiovisual Archives 20%
• Visual Features 20%
• Audio Features 20%
• Different Datasets 20%
• Training Model 20%
• Human Performance 20%
• Feature Combination 20%
• Deep Learning Methods 20%
• Action Recognition 20%
• Prediction Model 20%
• Encoder 20%
• State-of-the-art Models 20%
• Overall Architecture 20%
• Agnostic 20%
• Saliency Map 20%
• Multiple Hierarchies 20%
• 3D Convolution 20%
• Encoder-decoder Architecture 20%
• Fully Convolutional 20%
• Convolutional Encoder 20%
• Bilinear Interpolation 20%