Face landmark detection in Computer Vision - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load an image using OpenCV.

Computer Vision

import cv2
image = cv2.[1]('face.jpg')

Drag options to blanks, or click blank then click option'

Aimread

Bimshow

Cimwrite

Dresize

Attempts:

3 left

💡 Hint

Common Mistakes

Using cv2.imshow instead of cv2.imread will cause an error because imshow is for displaying images.

Using cv2.imwrite tries to save an image, not load it.

The function cv2.imread loads an image from a file into memory.

2fill in blank

medium

Complete the code to convert the image to grayscale for face landmark detection.

Computer Vision

gray = cv2.[1](image, cv2.COLOR_BGR2GRAY)

Drag options to blanks, or click blank then click option'

Aresize

BcvtColor

Cthreshold

Dblur

Attempts:

3 left

💡 Hint

Common Mistakes

Using cv2.resize changes image size, not color.

Using cv2.threshold applies a binary filter, not color conversion.

cv2.cvtColor converts images between color spaces, here from BGR to grayscale.

3fill in blank

hard

Fix the error in loading the pre-trained face landmark model.

Computer Vision

import dlib
predictor = dlib.shape_predictor('[1]')

Drag options to blanks, or click blank then click option'

Ashape_predictor_68_face_landmarks.dat

Bface_detector.xml

Clandmark_model.pkl

Dface_landmark.dat

Attempts:

3 left

💡 Hint

Common Mistakes

Using a file with wrong extension or name will cause file not found or loading errors.

Using face_detector.xml is for face detection, not landmarks.

The correct filename for the dlib 68-point face landmark predictor is shape_predictor_68_face_landmarks.dat.

4fill in blank

hard

Fill both blanks to detect faces and predict landmarks on the grayscale image.

Computer Vision

detector = dlib.get_frontal_face_detector()
faces = detector([1])
for face in faces:
    landmarks = predictor([2], face)

Drag options to blanks, or click blank then click option'

Agray

Bimage

Cgray, 0

Dimage, 0

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the color image to detector causes slower or incorrect detection.

Not passing the second argument 0 to predictor causes errors.

Face detection runs on the grayscale image. The predictor needs the grayscale image and a face rectangle with a second argument 0.

5fill in blank

hard

Fill all three blanks to extract landmark points and store them in a list.

Computer Vision

landmark_points = []
for n in range(0, [1]):
    x = landmarks.part(n).[2]
    y = landmarks.part(n).[3]
    landmark_points.append((x, y))

Drag options to blanks, or click blank then click option'

A68

Bx

Cy

D36

Attempts:

3 left

💡 Hint

Common Mistakes

Using 36 points extracts only eye landmarks, not full face.

Using .part(n).X or .part(n).Y (uppercase) causes attribute errors.

The dlib 68-point model has 68 landmarks indexed 0 to 67. Each point has x and y coordinates accessed by .x and .y.