Skip to content

Latest commit

 

History

History
35 lines (27 loc) · 1.92 KB

README.md

File metadata and controls

35 lines (27 loc) · 1.92 KB
Raw

Live-AI-Smile-Detector

An AI application that can detect the smile on the face(s) in real time i-e from a webcam or even a captured video.

Requirements

Apart from installing Python, you need to install the OpenCV and you're good to go.

The Classifiers

The "haarcascade_frontalface_default" is a pre-trained classifier that is used to detect faces in a frame. Also the "haarcascade_smile" is a pre-trained classifier that works the same way and is used to detect smiles on the faces.

The Code Flow

  • Import the most important OpenCv

  • Initialized the face and the smile detection classifiers and save it in the face_detector, smile_detector respectively.

  • Capture the webcam frames in the webcam

  • Now loop until the webcam is stopped and this continues

    • Read the points from the webcam
    • If something odd occurs, break out of the loop
    • Otherwise, convert the frame to Black and White (for opimization)
    • Detect faces in the frame, the Black and White frame
    • Draw the rectangle over the face(s) in the frame
    • Slicing the frame, the_face = frame[y:y + h, x:x + w], here we're cropping the whole frame into the sub-image in which we have only the faces.
    • Now convert that face' frame into Black and White
    • Detect the smiles within the faces' frame, which we converted in previous step
    • If want to have a rectangle over the smile you can
    • Now label "smiling" the face, if it detects the smile on the faces
    • Show the frame titling the "Live AI Smile Detector "
    • The cv2.waitkey(1) waits 1 milli second and then press the key to change the frame. If there is no argument you will need to press the key manually to change from frame to frame
    • At last release everything so that the operating system will come to know to free up the spaces

  • And that's pretty much it

License

This project is licensed under the MIT Open Source License.