Human Pose Estimation Data Annotation and Evaluation
Worked on human pose estimation using HRNet, focusing on annotation and evaluation on COCO and MPII datasets. Assisted in the preparation and validation of image data for fine-tuning and model evaluation. Handled high-precision multi-scale feature fusion and ensured accurate key point labeling for training. • Labeled human joints and key features for pose estimation tasks. • Used PyTorch for dataset processing and augmentation. • Evaluated accuracy and consistency of annotation for model input. • Coordinated with team to ensure dataset quality and compliance.