Efficient Stereo Matching Using Swin Transformer and Multilevel Feature Consistency in Autonomous Mobile Systems

Keyphrases

• Autonomous Mobile Systems 100%
• Coarse-to-fine Method 33%
• Computational Time Complexity 33%
• Contextual Information 33%
• Convolutional Neural Network 100%
• Cost Aggregation 33%
• Feature Consistency 100%
• Feature Extraction Tool 33%
• Full Use 33%
• Hierarchical Features 33%
• Image Region 33%
• KITTI Dataset 33%
• Low Computational 33%
• Matching Method 33%
• Memory Consumption 33%
• Multi-level Features 100%
• Multi-scale Feature Extraction 33%
• Network Transformers 33%
• Neural Network Method 66%
• Receptive Field 33%
• Scale Dependence 33%
• Scene Flow 33%
• Stereo Matching 100%
• Swin Transformer 100%
• Transformer 33%
• Unsupervised Method 33%

Computer Science

• Computational Time 33%
• Context Information 33%
• Convolutional Neural Network 100%
• Experimental Result 33%
• Feature Extraction 33%
• Inspired Method 33%
• Memory Consumption 33%
• Receptive Field 33%
• Stereo Matching 100%
• Swin Transformer 100%
• Time Complexity 33%
• Unsupervised Manner 33%

Engineering

• Computational Time 33%
• Context Information 33%
• Convolutional Neural Network 100%
• Experimental Result 33%
• Feature Extraction 33%
• Image Region 33%
• Matching Method 33%
• Mobile System 100%
• Multiscale 33%
• Receptive Field 33%
• Stereo Matching 100%