You need to implement a system that detects eye blink of a person in real-time mode. There is a useful guide that you can follow, describing how to implement eye blink detection from a given image. Your task is to continue the work applying the algorithm to detect eye blink of a certain duration. Also, be ready to answer questions about all the underlying mathematics.

The source of video could be a file or a stream from a web camera. On a video in the upper-left corner should be a caption that contains a number of blinks at the current time. If eyes are closed for more than two seconds, a caption “Alert!” must appear on the video. The colour of the caption should be red. You can use OpenCV, Dlib and whatever you find useful Minimum resolution of a video: 480p. There should be one person on a video. Height of a person’s face should be at least 1/3 of the height of the video in terms of pixels.

1. clone repository
2. open the .ipynb notebook in google colab or any other jupyter environment
3. make sure the files

blink_detection2.mp4
shape_predictor_68_face_landmarks.dat

are in the same working directory
4. run the notebook
5. download and play the file output.mp4

My apologies but my laptop got broken and is currently in Kazan for repairs. I used someone else's laptop and that is why it shows this person as an extra contributor