opencv-python

Reading images

img = cv2.imread("path/to/image/file.jpg")
# img is a numpy array
 
img = cv2.imread("path/to/image/file.jpg", 0)
# 0 => grayscale, 1 => bgr, -1 => preserve alpha values

Creating images

cv2.imwrite("path/to/file.jpg", old_img)

Resizing images

cv2.resize(image, (width, height))

Detecting face

To detect face, we need some models that will be used by the cv2 to process the image and detect face.

For example, haarcascade_frontalface_default.xml is a model

face_cascade = cv2.cascadeClassifier("haarcascade_frontalface_default.xml")
results = face_cascade.detectMultiScale(img, scaleFactor=1.05, minNeighbors=5)

results is an 2d array of detected positions.

[
	[x, y, width, height]
]

Drawing rectangle

This positions can be used to draw rectangle around faces

for x, y, w, h in results:
	img = cv2.rectangle(img, (x, y), (x + w, y + h), 3)

Showing image

cv2.imshow("Image", img)
cv2.waitKey(0)  # wait for key '0' to press
cv2.destroyAllWindows()  # after pressing '0' the image window will get closed

waitKey method wait for user key press on the image window. It takes parameter time in milliseconds to wait for the key. If it’s 0 then wait forever.

It returns the ascii value key pressed.

Video processing from webcam

video = cv2.VideoCapture(0) # 0 is the camera number
has_frame, frame = vide.read()
 
video.realase() # release the camera control

Blurring Images

img = cv.GaussianBlur(img, (21, 21), 0)

References

1. Cascade Classifier
2. opencv Haarcascades