-
Notifications
You must be signed in to change notification settings - Fork 3
Project Components
This interface that a user will use to interact with the system.
This is a custom identity-aware HTTP proxy that handles inbound traffic and relays it to internal services after successful authentication
The main functionality of this service is to expose endpoints to perform CRUD operations on Faces. It relies on other services such as the facecomparator and imageprocessor to generate the fingerprint in order to save/update a face.
This service handles all the image-related processing processes. It exposes three endpoints, each listed with its own features below
- /preprocess - Processes the video footage sent from the frontend frame by frame and sends the faces detected along with their respective timestamps
- /fingerprint - Generates the fingerprint for a face using the fingerprinter
- /coordinates - Finds the coordinates along with face data for a given frame from the front end. Used for the live detection feature.
This service receives video footage and removes the duplicate frames in it and returns it. It is used to reduce the file size and processing time for the system.
The sole task of this model is to extract all the face coordinates from the given frame.
Fingerprinter is a model completely written from scratch by the authors which is specific only for this system. As this is a deep learning research project, authors had to use trial and error methods while developing it. It generates a fingerprint for any given image.
This is a machine learning model, this model’s job is to match the fingerprints given by the fingerprinting model to the fingerprints on the database.