Who's There?

A facial recognition service developed in Python leveraging OpenCV to be deployed onto a security camera and identify authorized visitors

• Utilizes Dialogflow (API.AI) to process requests from Google Assistant enabled devices and interface with the facial recognition service backend
• Used the Frontal Face Haar Cascade for facial detection to crop out the faces while training the model
• Used OpenCV's Local Binary Patterns Histograms Algorithm for the Face Recognizer Class
• Includes user registration feature that detects faces and generates a dataset from a video source, using it to train the LBPH facial recognition algorithm
• Uses the Flask microframework web service to host the service

Supported Intents

The following intents are supported from Dialogflow along with the parameters passed onto the web service

Add User Intent

Resisters new user to the database and retrains model using new image dataset

• given-name: @sys.given-name (Required Parameter)
• num-samples: @sys.number-integer

Update User Intent

Updates model using new image dataset

• given-name: @sys.given-name (Required Parameter)
• num-samples: @sys.number-integer

Identify User Intent

Analyzes face from video feed to perform prediction on whether the user is authorized