A methodology for semantic action recognition based on pose and human-object interaction in avocado harvesting processes

Keyphrases

• Avocado 100%
• Harvesting System 100%
• Recognition-based 100%
• Human-object Interaction 100%
• Field Workers 100%
• Human Action Recognition 100%
• Feature Vector 66%
• Intelligent Systems 66%
• Pose Feature 66%
• Object Pose 66%
• Action Categories 66%
• Human Action 66%
• Combined Object 66%
• Chilean 33%
• Principal Coordinate Analysis (PCoA) 33%
• Increased Productivity 33%
• Support Vector Machine 33%
• K-nearest 33%
• Decision Tree 33%
• Agricultural Industry 33%
• Estimation Method 33%
• Dimensionality Reduction 33%
• Classification Performance 33%
• Faster R-CNN 33%
• Machine Decisions 33%
• Production Increase 33%
• Region-based Convolutional Neural Network (R-CNN) 33%
• Convolutional Object Detector 33%
• Field Tools 33%
• Still Images 33%
• Bagged Tree 33%
• Skeleton Joints 33%
• Semantic Information 33%
• InceptionResNetV2 33%
• Human Skeleton 33%
• Single Feature 33%
• Inference Time 33%
• Neighbor-joining Tree 33%
• 2D Pose Estimation 33%
• Field Vehicle 33%
• Convolutional 33%
• Human Labor 33%
• OpenPose 33%
• Action Classification 33%
• Accuracy Performance 33%

Computer Science

• Convolutional Neural Network 100%
• Feature Vector 100%
• Object Pose 100%
• Action Recognition 100%
• Principal Components 50%
• Decision Trees 50%
• Robot 50%
• Component Analysis 50%
• Support Vector Machine 50%
• Classification Performance 50%
• Single Feature 50%
• Average Accuracy 50%
• Object Detection 50%
• Pose Estimation 50%
• Estimation Method 50%