20_mystery-men

Sam Frey (github: samfrey99)

Alvaro Albero (github: albero94)

Description

We have created a smart door application that uses face recognition. The basic description of the service is, we have a cloud application that contains the list and images of people that are allowed into the house. We have a mobile application that is used as an administrator and can add and delete people from the list and operate the door lock directly. We have an IoT system with a Raspberry PI, a camera and a lock actuator that will take a picture of the person trying to get into the house, send it to the cloud service, and operate the door if there is a match in the face.

• Face Recognition Server contains the cloud service
• Android contains the mobile application
• PI contains the Raspberry PI code to manage the actuators
• Deployment contans the files to deploy the project
• A short demo video can be found here.

Deployment

If you want to use our application, you need the following components. A hosting environment to deploy the .NET server, an Azure Face API subscription (basic is free), an android device to administrate the system, a camera, a door lock, and a Raspberry PI to operate both of them.

We have included a Deployment folder where you can find the three projects (Cloud, Web and PI).

To deploy the cloud application

• You need to copy the files in FaceRecognitionServer to the hosting provider of your choice, we have used Azure App Services
• You need to register to Azure Face API to get the API keys, then include them as environment variables named AZURE_FACE_SUBSCRIPTION_KEY and AZURE_FACE_ENDPOINT
• Your cloud service is ready to use

To deploy the mobile application

• Download the APK and install on an Android Device, or build the project in Android Studio and run in an emulator.

To deploy the IoT component

• Download controller.py and run on a Raspberry Pi.
• A wiring example can be found here