Video writing in Computer Vision - ML Experiment: Train & Evaluate

Experiment - Video writing

Problem:You want to save a sequence of images as a video file using OpenCV in Python.

Current Metrics:The current code writes a video but the output video is corrupted or does not play properly.

Issue:The video writer parameters (codec, frame size, fps) are not set correctly, causing the video file to be unreadable.

Your Task

Fix the video writing code so that the saved video plays correctly with the expected frame size and frame rate.

Use OpenCV (cv2) for video writing.

Keep the frame size and fps consistent with the input images.

Do not change the source images.

Hint 1

Hint 2

Hint 3

Solution

Computer Vision

import cv2
import numpy as np

# Create dummy images (e.g., 100 frames of 640x480 with random colors)
frame_width, frame_height = 640, 480
num_frames = 100
frames = [np.random.randint(0, 256, (frame_height, frame_width, 3), dtype=np.uint8) for _ in range(num_frames)]

# Define video writer parameters
fourcc = cv2.VideoWriter_fourcc(*'mp4v')  # Codec for mp4
fps = 20
out = cv2.VideoWriter('output_video.mp4', fourcc, fps, (frame_width, frame_height))

for frame in frames:
    out.write(frame)  # Write each frame

out.release()  # Release the video writer
print('Video writing completed successfully.')

Set the codec to 'mp4v' which is widely supported for mp4 files.

Ensured the frame size matches exactly the size of the input images (640x480).

Set fps to 20 for smooth playback.

Used a proper VideoWriter object with correct parameters.

Results Interpretation

Before: Video file was corrupted or would not play due to incorrect codec or frame size.
After: Video plays smoothly with correct resolution and frame rate.

Setting the correct codec, frame size, and fps is essential for successful video writing. Mismatched parameters cause corrupted or unplayable video files.

Bonus Experiment

Try writing the video in grayscale instead of color.

💡 Hint

Convert each frame to grayscale using cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) and adjust the VideoWriter parameters accordingly (set isColor=False).

Practice

1. What is the main purpose of using cv2.VideoWriter in computer vision?

easy

A. To save a series of images as a video file

B. To read video files frame by frame

C. To apply filters on images

D. To display images on screen

Solution

1. Step 1: Understand the function of cv2.VideoWriter

   cv2.VideoWriter is designed to save multiple images as a continuous video file.

2. Step 2: Differentiate from other OpenCV functions

   Functions like cv2.VideoCapture read videos, while VideoWriter writes videos.

3. Final Answer:

   To save a series of images as a video file -> Option A

4. Quick Check:

   Video writing = save images as video [OK]

Hint: VideoWriter saves frames into video files [OK]

Common Mistakes:

• Confusing VideoWriter with VideoCapture
• Thinking VideoWriter displays images
• Assuming VideoWriter applies image filters

2. Which of the following is the correct way to create a cv2.VideoWriter object for a 640x480 video at 30 fps using the 'XVID' codec?

easy

A. cv2.VideoWriter('output.avi', cv2.VideoWriter_fourcc(*'XVID'), 30, (640, 480))

B. cv2.VideoWriter('output.avi', 'XVID', 30, (640, 480))

C. cv2.VideoWriter('output.avi', cv2.VideoCapture_fourcc('XVID'), 30, (640, 480))

D. cv2.VideoWriter('output.avi', cv2.VideoWriter_fourcc('X', 'V', 'I', 'D'), 30, 640, 480)

Solution

1. Step 1: Use cv2.VideoWriter_fourcc correctly

   The codec must be passed as cv2.VideoWriter_fourcc(*'XVID') to get the fourcc code.

2. Step 2: Check the parameters order and types

   Parameters are filename, fourcc, fps, and frame size as a tuple (width, height).

3. Final Answer:

   cv2.VideoWriter('output.avi', cv2.VideoWriter_fourcc(*'XVID'), 30, (640, 480)) -> Option A

4. Quick Check:

   Correct codec and frame size syntax = cv2.VideoWriter('output.avi', cv2.VideoWriter_fourcc(*'XVID'), 30, (640, 480)) [OK]

Hint: Use cv2.VideoWriter_fourcc(*'XVID') for codec [OK]

Common Mistakes:

• Passing codec as string directly
• Using cv2.VideoCapture_fourcc instead of VideoWriter_fourcc
• Incorrect frame size format (not a tuple)

3. What will be the output of this code snippet?

import cv2
fourcc = cv2.VideoWriter_fourcc(*'MJPG')
out = cv2.VideoWriter('test.avi', fourcc, 20.0, (320,240))
frame = 255 * np.ones((240,320,3), dtype=np.uint8)
out.write(frame)
out.release()
print('Video saved')

medium

A. Nothing happens, no file is created

B. A video file named 'test.avi' with a white frame is saved and 'Video saved' is printed

C. RuntimeError because frame size does not match

D. NameError due to missing import of numpy

Solution

1. Step 1: Check imports

   The code uses np.ones but numpy (np) is not imported, causing a NameError.

2. Step 2: Understand error impact

   Without importing numpy, the code will fail before writing or saving video.

3. Final Answer:

   NameError due to missing import of numpy -> Option D

4. Quick Check:

   Missing numpy import = NameError [OK]

Hint: Always import numpy when using np [OK]

Common Mistakes:

• Assuming code runs without numpy import
• Ignoring frame size correctness
• Thinking print means success

4. Identify the error in this video writing code and select the fix:

import cv2
fourcc = cv2.VideoWriter_fourcc(*'XVID')
out = cv2.VideoWriter('output.avi', fourcc, 25, (640, 480))
frame = cv2.imread('image.jpg')
out.write(frame)
out.release()

medium

A. Fix: Change codec to 'MJPG'

B. Fix: Resize frame to (640, 480) before writing

C. Fix: Use grayscale image instead of color

D. Fix: Increase fps to 30

Solution

1. Step 1: Check frame size vs VideoWriter size

   VideoWriter expects frames of size (640, 480). If frame size differs, write() fails or produces errors.

2. Step 2: Correct frame size by resizing

   Resize the loaded image to (640, 480) before writing to match VideoWriter settings.

3. Final Answer:

   Fix: Resize frame to (640, 480) before writing -> Option B

4. Quick Check:

   Frame size must match VideoWriter size [OK]

Hint: Match frame size to VideoWriter size [OK]

Common Mistakes:

• Ignoring frame size mismatch
• Changing codec unnecessarily
• Changing fps unrelated to error

5. You want to create a video from a list of images with different sizes. Which approach correctly handles this situation?

hard

A. Use VideoWriter without resizing; it auto-adjusts frame size

B. Write images directly without VideoWriter; just save as separate files

C. Resize all images to the same size before writing to VideoWriter

D. Change codec to 'X264' to support variable frame sizes

Solution

1. Step 1: Understand VideoWriter frame size requirement

   VideoWriter requires all frames to have the same size as specified at creation.

2. Step 2: Handle varying image sizes

   Resize all images to a consistent size before writing to avoid errors and ensure smooth video.

3. Final Answer:

   Resize all images to the same size before writing to VideoWriter -> Option C

4. Quick Check:

   Consistent frame size needed for video writing [OK]

Hint: Resize images to uniform size before video writing [OK]

Common Mistakes:

• Assuming VideoWriter auto-resizes frames
• Trying to write images without VideoWriter
• Changing codec to fix size issues