Here is a template project to reuse for production ready applications to use Deep Learning models focusing on Face, Age, Gender detection models.
At this stage, only OpenVINO has been integrated.
OpenVINO(TM) Version: 2019.2.242
You can install OpenVINO(TM) by following the instructions published online documentation.
Before running this Python application:
- Set Environment Variables on the current workspace:
source /opt/intel/openvino/bin/setupvars.sh
OR
- Set Environment Variables System Wide
Copy files/intel-openvino.sh
& files/intel-openvino.conf
file as shown below:
sudo cp files/intel-openvino.sh /etc/profile.d/
sudo cp files/intel-openvino.conf /etc/ld.so.conf.d/
sudo reboot
git clone https://github.com/odundar/face_detection.git
If all setup completed successfully, you can use the default configurations to give a start for face detection application.
python3 face_detection_openvino.py config/config.json
app
folder includes apps ready to run for face, age, gender detection applications.
service modules stored here which are to be deployed in docker microservices
config
folder includes app and service default configurations to be used as template.
detection
folder contains the modules and classes to reuse for inference application development.
detection\
detection_base_ov.py\
InferenceConfig
InferenceBase
age_gender_detection_ov.py\
AgeGenderDetectionTypes
AgeGenderDetection
AgeGenderConfig
MTCNNAgeGenderDetection
MTCNNAgeGenderConfig
face_detection_ov.py\
FaceDetectionModelTypes
FaceDetectionConfig
OpenMZooFaceDetection
MTCNNFaceDetectionConfig
MtCNNFaceDetection
Includes instructions to deploy face, age-gender detection services as docker services.
This folder includes system configuration files for OpenVINO(TM) Toolkit
Folder includes instructions how to fetch models, convert and use them.