Fast Video Instance Segmentation via Recurrent Encoder-Based Transformers

Keyphrases

• Encoder 100%
• Transformer 100%
• Video Instance Segmentation 100%
• Spatio-temporal Features 50%
• Inference Speed 50%
• Segmentation Framework 33%
• Segmentation Performance 33%
• Popular 16%
• Computationally Expensive 16%
• YouTube Videos 16%
• Transformer-based 16%
• Overfitting 16%
• Attention-based 16%
• Capture Target 16%
• CNN Encoder 16%
• Attention-based Encoder-decoder 16%
• Custom CNN 16%
• Target Appearance 16%
• Knowledge Distillation 16%

Computer Science

• Instance Segmentation 100%
• Segmentation Performance 33%
• YouTube 16%
• Knowledge Distillation 16%