OpenPose overview in Computer Vision - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load an OpenPose model using the OpenPose Python API.

Computer Vision

from openpose import pyopenpose as op
params = dict()
params["model_folder"] = "models/"
opWrapper = op.WrapperPython()
opWrapper.configure(params)
opWrapper.[1]()

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

Arun

Binit

Cload

Dstart

Attempts:

3 left

💡 Hint

Common Mistakes

Using run() instead of start(), which does not exist in the API.

Using load() which is not a method of the wrapper.

Using init() which is not the correct method name.

The start() method initializes and starts the OpenPose wrapper to process images.

2fill in blank

medium

Complete the code to process an image and get the pose keypoints.

Computer Vision

datum = op.Datum()
datum.cvInputData = image
opWrapper.emplaceAndPop([[1]])
keypoints = datum.poseKeypoints

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

AopWrapper

Bdatum

Cparams

Dimage

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the image directly instead of the datum object.

Passing the wrapper or params which are not valid inputs here.

The emplaceAndPop method takes a list of Datum objects to process. We pass datum which contains the input image.

3fill in blank

hard

Fix the error in the code to correctly extract the number of detected people from pose keypoints.

Computer Vision

num_people = datum.poseKeypoints.[1] if datum.poseKeypoints is not None else 0

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

Ashape[0]

Blen()

Csize

Dcount

Attempts:

3 left

💡 Hint

Common Mistakes

Using len() which causes an error because poseKeypoints is a NumPy array, not a list.

Using size which gives total elements, not number of people.

Using count which is not a valid attribute.

The poseKeypoints is a NumPy array where the first dimension is the number of people detected. Using shape[0] gives the count.

4fill in blank

hard

Fill both blanks to create a dictionary of body part names and their 2D coordinates for the first detected person.

Computer Vision

body_parts = {name: (datum.poseKeypoints[0, idx, [1]], datum.poseKeypoints[0, idx, [2]]) for idx, name in enumerate(body_part_names)}

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

A0

B1

C2

D3

Attempts:

3 left

💡 Hint

Common Mistakes

Swapping x and y indices.

Using index 2 which is confidence, not coordinate.

The 2D coordinates are stored in the second and first indices of the last dimension: 1 for x and 0 for y.

5fill in blank

hard

Fill all three blanks to filter detected people with confidence above 0.5 and create a list of their nose coordinates.

Computer Vision

nose_coords = [datum.poseKeypoints[i, nose_idx, [1]] for i in range(datum.poseKeypoints.shape[0]) if datum.poseKeypoints[i, nose_idx, [2]] > [3]]

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

A0

B2

C0.5

D1

Attempts:

3 left

💡 Hint

Common Mistakes

Using wrong indices for coordinates or confidence.

Using a confidence threshold other than 0.5.

The x coordinate is at index 1, confidence at index 2, and the threshold is 0.5.