Learning single and multi-scene camera pose regression with transformer encoders

Keyphrases

• Camera Pose Regression 100%
• Multi-scene 100%
• Transformer Encoder 100%
• Regress 50%
• Camera Pose 50%
• Decoder 25%
• Encoder 25%
• Feature Matching 25%
• Transformer 25%
• Publicly Available 25%
• 3D Scene 25%
• Query Image 25%
• 3D Coordinates 25%
• Contemporary State 25%
• Random Sample Consensus (RANSAC) 25%
• Spatially Variable 25%
• Regression Approach 25%
• Latent Representation 25%
• Sub-meter 25%
• Absolute Camera Pose 25%
• Multiple Scenes 25%
• Transformer-based 25%
• Pose Regression 25%
• Outdoor Scenes 25%
• Scene Modeling 25%
• Camera Intrinsic Parameters 25%
• Localization Technique 25%
• Inference Time 25%
• Structured Scenes 25%

Computer Science

• 3d Scenes 100%
• localization algorithm 50%
• Average Accuracy 50%
• Deep Feature 50%
• Reference Image 50%