Multi-Modal Fusion of Event and RGB for Monocular Depth Estimation Using a Unified Transformer-based Architecture

Keyphrases

• Transformer 100%
• Depth Estimation 100%
• Multimodal Fusion 100%
• Transformer Architecture 100%
• Pixel Level 66%
• High Temporal Resolution 33%
• Absolute Intensity 33%
• Recurrent Neural Network 33%
• Contextual Interaction 33%
• Relative Difference 33%
• Modal 33%
• Neural Network Method 33%
• Wide Dynamic Range 33%
• Source Code 33%
• Relative Depth 33%
• Proposed Architecture 33%
• Difference Threshold 33%
• Depth Prediction 33%
• Depth Threshold 33%
• Asynchronous Events 33%
• Event Dynamics 33%
• State-of-the-art Models 33%
• Autonomous Navigation 33%
• Mean Depth 33%
• Camera Vision 33%
• Robotic Navigation 33%
• Dynamic Vision Sensor 33%
• Input Modality 33%
• Self-attention 33%
• RGB Image 33%
• Motion Blur 33%
• Transformer Model 33%
• Scene Context 33%

Computer Science

• Depth Estimation 100%
• Recurrent Neural Network 25%
• Temporal Information 25%
• RGB Image 25%
• Temporal Resolution 25%
• Relative Difference 25%
• Autonomous Navigation 25%

Engineering

• Multi-Modal Fusion 100%
• Traditional Camera 50%
• Temporal Information 50%