Train a ML model to recognize my spouse and I when we get home, and then have it greet us with a fun fact. This involved writing a function to query the wikipedia API, and then sending the results to AWS polly to synthesize the text.
I also want this project to be as dynamic as possible, so that regardless of how many people the model is trained on it will have the same ability to uniquely identify and greet them without requiring any hard coding to do so.
Images: Add a directory in the images folder with the name of each person you want the model to recognize. Crop a photo of the face of the person, and place one single photo in the directory as a .jpg.
AWS Polly: In polly.py replace profile_name="personal-aws" with the name of your AWS access keys. On linux and mac these can be found in ~/.aws/credentials
Configure a video stream I've used the free android app, IP Webcam, and in main.py you can set the rtsp end point for streaming video to the greeter.