GrabCut

GrabCut

1. GrabCut

• Image segmentaion based algorithm
• The algorithm estimates the color distribution of the target object and that of the background using a Gaussian mixture model.
• Automatically distribute when user select rectangle.

• GrabCut code

cv2.grabCut(img, mask, rect, bgdModel, fgdModel, iterCount, mode=None) -> mask, bgdModel, fgdModel

• img : Input image
• mask : Input and Output mask
• rect : Roi area
• bgdModel : Temporary background model matrix
• fdgModel : Temporary foreground model matrix
• iterCount : Number of iteration
• mode : Mode constant

2. Moment based object detection

• Moment
  • Real number of expressing the form of image
  • It is calculated to correlation of specific function set
  • 
• Hu’s seven invariant moments
  • Seven moment value central moment put together nomalizaion under 3-dimension
  • It is unchanging to rotaion, movement, reflection, size of image

• Moment based object detection code

cv2.matchShapes(contour1, contour2, method, parameter) -> retval

• contour1 : First outline or grayscale image
• contour2 : Second outline or grayscale image
• method : Signify comparision method
• parameter : Not use
• retval : Two outline or distance of grayscale image

• Shape comparision function

• Moment based object detection example