Docker_Occlusion

A pose estimator that is more robust to face occlusions. The code includes a gRPC server and client example to perform the estimation in images.

DOWNLOAD THE ESTIMATOR MODEL HERE: https://drive.google.com/file/d/19JgO63Zhz0CJ3kdVFr5u4I0xWngc0vSz/view?usp=sharing

gRPC Server and Client

How to experiment

Download this repository. Open two terminals and go to the repository folder directory. On one of the terminals run python OcclusionPose_server.py. On the other run python OcclusionPose_client.py face_image.jpg (face_image will be your image name). It will output the pose and the image with the pose axis.

Generate protobuf files

If you desire to change the proto file run python -m grpc_tools.protoc -I. --python_out=. --grpc_python_out=. Occlusion_pose.proto This will generate the pb2 and pb2_grpc files.

Docker

Open a terminal and type docker run -p 8061:8061 -it --rm josecarlos714/dockerpose:occlusionpose. This will run a docker container which will deploy the pose estimation server. You can extract the pose using the OcclusionPose_client.py file (an example file) for a given image.

Docker Compose

Go to docker_pipeline folder and run docker-compose up to visualize pose estimation of face images in the images folder.