• AWS DeepLens v1.1
• python 3.7
• mxnet
• ubuntu

1. You can train any posture
2. Extract posture data using extract_pose.py
3. Train using mxnet_train.ipynb
4. Get 'DeepLens_pose-0500.params' and 'DeepLens_pose-symbol.json'

CMU-Perceptual-Computing-Lab/openpose

mxnet_Realtime_Multi-Person_Pose_Estimation

1. Convert openpose model to run on DeepLens
2. Get 'realtimePose.bin' and 'realtimePose.xml'

Reference : deeplens-inference-lambda-create

1. Use AWS Lambda

python deeplens_pose_live.py

python deeplens_pose_image.py

./view_deeplens_project_streams.sh

./view_deeplens_live_streams.sh