opencv/examples/coffeescript/camera-face-detection.coffee

Node Bindings to OpenCV

cv = require 'opencv'

# First we create a new VideoCapture object to get
# video from the camera (0 for default camera)
camera = new cv.VideoCapture(1)

# we create a window to display the Video frames
namedWindow = new cv.NamedWindow('Video',1)
color = [255,0,0]

# We set an interval to retrieve frames from the
# video source and we get the intervalId so we can
# stop the program from the video feed window.
intervalId = setInterval(()->

  # camera.read allows us to retrieve the current
  # frame to be displayed in the video window.
  camera.read((err, im) ->
    # We can check for errors and even break the
    # program execution if an error is detected here.
    console.log "The err ==>#{ err }" if err

    # There is no need to display the image width or Height
    # but I leave this here in case anyone needs to check them.
    #console.log("Width: #{im.width()}")
    #console.log("height: #{im.height()}")

    # Before working with the frame we need to check the image
    # is already available and has a width and height greater than 0,
    # otherwise it will fail when trying to do namedWindow.show()
    # and the image has width or height equal or less than 0.
    if im.width() > 0 and im.height() > 0
      im.detectObject('./haarcascades/haarcascade_frontalface_alt.xml', {}, (err, faces) ->
        for face in faces
          im.rectangle([face.x, face.y], [face.x + face.width, face.y + face.height], [0,255,0], 2)

        # We use the previously created namedWindow to display the
        # video frame to wich we applied the blur and filter.
        namedWindow.show(im)
      )

    # Finally we get the key pressed on the window to terminate
    # execution of the program.
    res = namedWindow.blockingWaitKey(0, 50)

    # In this case I terminate the program if any key is pressed.
    if res >= 0 then clearInterval(intervalId)
  )
, 150)