Facial Recognition
This facial recognition project is an integration of a joint work between different teams.
Modules
The project has four modules:
- Face Detection: Detects the largest face of a given image and returns its location.
- Face Aligner: Detects the important landmark points of a given face and returns an aligned face.
- Face Descriptor Extractor: Describes the features of a given aligned image so that it can recognize different images.
- DB: Saves and fetches information of given features to be used to recognize a certain person.
Cases
This project allows to perform three different cases:
- Validate: Compares a taken picture with the descriptors of a saved picture in the DB for a given ID. Returns true if the person is recognized as the same one from the DB.
- Recognize: Gives the x amount of people with the highest amount of similarity between all the people saved in the DB.
- Register: Saves the given person, with its picture, ID, and other relevant data into the DB.
Tools
- Dlib library: http://dlib.net/
- Open CV
- Alignment models downloaded from this repository.