Repository for master thesis 'Deep Learning-Based Human Pose Estimation from 3D Data'
\DDP - source code for re-implementation of Deep Depth Pose model in Keras
- launch DDP.py to run the model on ITOP dataset
- launch DDP_UBC.py to run the model on UBC3V dataset
\PBPE - source code for re-implementation of Point-Based Pose Estimation model in Keras, and implementation of our novel method Segmentation-Guided Pose Estimation
- parameters setup in config.py
- launch main.py to run the specified model on specified data
\trained_models - saved trained models for each dataset for testing
\sample_data - 200 samples from test data for each dataset
© Dana Škorvánková 2020