AI Facial Recognition web system

Facial recognition and facial detection web system in Python and Django.

Ready-to-use code with User Interface for facial recognition and facial detection. Upload images and let the AI find matches for people.

Upload individual or multiple images at the time.


style="color:#000000;">While facial recognition gives an identification or a match between a set of images, facial detection allows identifying if there is a human face (or more) within a picture.

With the following project you will be able to:

• Use it as it is for web facial recognition and facial detection


• Upload single images or multiple images to build a known faces library


• Find a match by facial recognition for a single image or multiple files


• Submit an image for facial detection


• Browse the results as Web client


• Access to the database and source code


• Use the recognition class for other projects

Includes pre-trained data (pre-trained data from public libraries*, you can also use your own training dataset or another public one)

External Sources

The model was trained using the following dataset: https://ibug.doc.ic.ac.uk/resources/facial-point-annotations/. (please check their website for changes in License for the landmarks; additionally, you can use a Tensorflow MIT dataset for training).

For this system, we will use the face_detector to find the faces.  Alternatively, you can use your own datasets and replace those files.

How to Use:

First, copy the two .dat files from the "Part2" zip folder inside the Part1 zip folder (where the manage.py file is).

Install Python3 and Django in your server/localhost

Copy the folder in your decided location (or server’s html folder).

In the terminal cd inside the folder with the manage.py file.

Type the followings:

> python3 manage.py makemigrations

> python3 manage.py migrate

> python3 manage.py runserver

Limitations

• The detections correspond to statistical probabilities, hence the accuracy is not 100% guaranteed.
  


• The image quality and complexity interferes with the accuracy of the prediction.