Writing/saving images in Computer Vision - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

What is the output of this OpenCV image saving code?

Consider the following Python code using OpenCV to save an image. What will be the output or result after running this code?

Computer Vision

import cv2
import numpy as np

img = np.zeros((100, 100, 3), dtype=np.uint8)
cv2.imwrite('test_image.png', img)
print(cv2.imread('test_image.png').shape)

A (100, 100, 3)

B (100, 100)

C None

D Error: File not found

💡 Hint

Remember that OpenCV reads color images as 3-channel arrays by default.

The code creates a black image of size 100x100 with 3 color channels, saves it as 'test_image.png', then reads it back. The shape of the read image is (100, 100, 3).

❓ Model Choice

intermediate

Which image format is best for saving photos with lossless quality?

You want to save images without losing any quality for further machine learning processing. Which image format should you choose?

A JPEG

B BMP

C GIF

D PNG

💡 Hint

Consider which formats compress images without losing data.

PNG is a lossless compression format, preserving all image data, unlike JPEG which uses lossy compression.

❓ Hyperparameter

advanced

What does the 'quality' parameter control when saving JPEG images with OpenCV?

In OpenCV's cv2.imwrite function, you can specify parameters like cv2.IMWRITE_JPEG_QUALITY. What does this parameter control?

A The compression level affecting image quality and file size

B The color depth of the saved image

C The image resolution in pixels

D The file permissions of the saved image

💡 Hint

Think about how JPEG compression works.

The 'quality' parameter sets the compression level; higher values mean better quality and larger file size.

🔧 Debug

advanced

Why does this code fail to save the image correctly?

Look at this code snippet: import cv2 import numpy as np img = np.zeros((50, 50), dtype=np.uint8) cv2.imwrite('output.jpg', img) Why might the saved image appear as a black and white image instead of color?

A The numpy array dtype must be float32 for color images

B The file extension '.jpg' is incorrect for color images

C The image array has only one channel, so it saves as grayscale

D cv2.imwrite requires images to be in RGB order, not BGR

💡 Hint

Check the shape of the numpy array representing the image.

The image array has shape (50, 50) with no color channels, so OpenCV saves it as a grayscale image.

🧠 Conceptual

expert

What is the main difference between cv2.imwrite and PIL.Image.save when saving images?

Both OpenCV's cv2.imwrite and PIL's Image.save can save images. What is a key difference in how they handle image data before saving?

A OpenCV only saves grayscale images, PIL saves color images

B OpenCV expects images in BGR format, PIL expects RGB format

C PIL compresses images automatically, OpenCV does not

D OpenCV saves images as numpy arrays, PIL saves as tensors

💡 Hint

Think about the color channel order each library uses.

OpenCV uses BGR channel order by default, while PIL uses RGB, so images may appear with swapped colors if not converted.