Why 3D understanding enables robotics and AR in Computer Vision - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a 3D point cloud from depth data.

Computer Vision

point_cloud = depth_image.[1]()

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

Aflatten

Bmean

Csum

Dreshape

Attempts:

3 left

💡 Hint

Common Mistakes

Using sum or mean instead of reshape changes the data incorrectly.

We use reshape to convert the depth image into a 3D point cloud format.

2fill in blank

medium

Complete the code to estimate the camera pose using 3D points.

Computer Vision

pose = estimate_pose([1], camera_intrinsics)

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

Apoint_cloud_3d

Bimage_2d_points

Ccolor_image

Ddepth_map

Attempts:

3 left

💡 Hint

Common Mistakes

Using 2D points or color images instead of 3D points.

Estimating camera pose requires 3D points, so we use point_cloud_3d.

3fill in blank

hard

Fix the error in the code to transform 3D points using a rotation matrix.

Computer Vision

transformed_points = rotation_matrix @ [1]

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

Apoints_3d.T

Bpoints_3d

Cpoints_3d.flatten()

Dpoints_3d.sum(axis=1)

Attempts:

3 left

💡 Hint

Common Mistakes

Not transposing points causes dimension mismatch errors.

The points must be transposed to match matrix multiplication dimensions.

4fill in blank

hard

Fill both blanks to filter 3D points within a certain distance.

Computer Vision

filtered_points = {p for p in points if p.[1] [2] max_distance}

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

Adistance()

B<

C>

Dnorm()

Attempts:

3 left

💡 Hint

Common Mistakes

Using incorrect method names or wrong comparison operators.

We use norm() to get the distance of point p and keep points with distance less than max_distance.

5fill in blank

hard

Fill all three blanks to create a dictionary of 3D points and their colors for AR rendering.

Computer Vision

point_color_map = [1]: [2] for [3], [2] in zip(points_3d, colors) if [2] is not None

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

Ap

Bc

Attempts:

3 left

💡 Hint

Common Mistakes

Mixing up variable names or forgetting to check for None colors.

We map each 3D point p to its color c, iterating over pairs from zip(points_3d, colors).