Learning Semantic Correspondence Exploiting an Object-Level Prior

Engineering

• Benchmark 50%
• Convolutional Neural Network 100%
• Deformation 50%
• Experimental Result 50%
• Flow Field 50%
• Geometric Transformation 50%
• Good Compromise 50%
• Neural Network Architecture 50%
• Point Correspondence 50%

Computer Science

• Benchmark 50%
• Convolutional Neural Network 100%
• Convolutional Neural Network 100%
• Experimental Result 50%
• Foreground Mask 50%
• Geometric Transformation 50%
• Specific Detail 50%
• Training Data 50%