Self-Supervised Learning via Multi-Transformation Classification for Action Recognition

Keyphrases

• Action Recognition 100%
• Human Action 100%
• Downstream Task 100%
• Self-supervised Learning 100%
• Learning Model 50%
• Learning Methods 50%
• Extracting Features 50%
• Video Clips 50%
• Contextual Information 50%
• 3D Convolutional Neural Network (3D CNN) 50%
• Visual Representation 50%
• Backbone Network 50%
• Training Model 50%
• HMDB51 50%
• UCF-101 50%
• ResNet18 50%
• 3D ResNet 50%
• Noise Addition 50%
• Pseudo Label 50%
• Pretext Task 50%
• Color Switch 50%
• Spatio-temporal Representation 50%
• Video Representation Learning 50%
• Video Transformation 50%

Computer Science

• Action Recognition 100%
• Self-Supervised Learning 100%
• Experimental Result 50%
• Annotation 50%
• Contextual Information 50%
• Backbone Network 50%
• Visual Representation 50%
• Representation Learning 50%
• 3D Convolutional Neural Networks 50%
• Pre-Trained Model 50%
• Video Representation 50%
• Noise Addition 50%