TransformerFusion: Monocular RGB Scene Reconstruction using Transformers

Keyphrases

• Monocular RGB 100%
• Transformer 100%
• Scene Reconstruction 100%
• 3D Reconstruction 66%
• Coarse-to-fine 66%
• Reconstruction Approach 66%
• Transformer Architecture 33%
• Image Frame 33%
• Reconstruction Problem 33%
• Video Sequences 33%
• Fully Convolutional 33%
• Surface Reconstruction 33%
• Transformer Network 33%
• 3D Location 33%
• Transformer-based 33%
• Video Frames 33%
• Low Memory 33%
• 3D Features 33%
• Multi-view Stereo 33%
• Occupancy Prediction 33%
• RGB-D Video 33%
• Memory Storage 33%
• 3D Scene Representation 33%
• Volumetric Features 33%
• Recurrent Network 33%
• Stereo Depth Estimation 33%
• Relevant Images 33%
• Fine Levels 33%
• Depth Estimation Method 33%
• Sequence Fusion 33%

Computer Science

• 3d Scenes 100%
• Surface Reconstruction 50%
• 3d Reconstruction 50%
• Video Sequences 50%
• Long Short-Term Memory Network 50%
• Depth Estimation 50%
• Recurrent Network 50%
• Estimation Method 50%