Learning to synthesize 3D indoor scenes from monocular images

Computer Science

• Convolutional Neural Network 100%
• Long Short-Term Memory Network 100%
• Experimental Result 50%
• Object Detection 50%
• Depth Information 50%
• Art Performance 50%
• Semantic Label 50%
• 3d Scenes 50%
• Geometric Information 50%
• Scene Understanding 50%
• RGB Image 50%
• Long Short-Term Memory Networks 50%

Keyphrases

• Convolutional Long Short-term Memory (ConvLSTM) 100%
• 3D Indoor Scene 100%
• Regression Convolutional Neural Network 66%
• Geometric Loss 33%
• Indoor Scene Understanding 33%
• Object Arrangement 33%
• Monocular RGB Image 33%
• 2D Segmentation 33%
• 3D Inference 33%
• Network Benefits 33%
• Joint Inference 33%