Camera face recognition and directionality tracking + Website and Mobile App for data entry

The project consists in a working camera with face recognition and people tracking directionality embedded (edge computing) from a top view position. The camera will be fisheye camera and will be installed on a frame door (see the attached image).
In addition, we want to connect the camera with a website and mobile app for data entry for further ML training.

Requirements:
1. Include entire hardware solution. ESP-EYE or similar board (small board, not raspberry pi or nvidia jetson).
2. Using fisheye camera (board compatible) from a top view position as it will be installed on a frame door.
3. Edge Computing with Person detection and face recognition with high accuracy (more than 95%) and more than 8fps.
4. People Tracking with directionality.
5. People counting.
6. C/C++ for better performance.
7. The camera will be connected with a Mobile App that collects user images with a mobile camera and automatically train/retrain a pre-trained ML model + Log in functionality
8. The camera will be connected with a Web App for entry users data and images to train a ML model + Log in functionality
9. All the code commented.
10. Log file with the application outputs

For more details, see the attached document.