Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/ilirosmanaj/detect_kermit/llms.txt

Use this file to discover all available pages before exploring further.

convert_vid2image.py reads a video file using OpenCV and writes one JPEG frame per second to disk, producing the raw images needed for dataset preparation. These frames form the foundation of the Kermit detection training and test sets — episodes 1 and 2 supply training images, while episode 3 supplies the validation set.

​
Source code

helpers/convert_vid2image.py
import cv2

count = 0
cap = cv2.VideoCapture('data/video.avi')
ret, image = cap.read()

print('Starting getting frames from video...')
while ret:
    cap.set(cv2.CAP_PROP_POS_MSEC, (count * 1000))
    ret, image = cap.read()
    if not ret:
        break

    cv2.imwrite( 'data/videoframe/' + "frame%d.jpg" % count, image)
    count = count + 1

​
How to use

  1. Place your video file at data/video.avi (or edit the path string in the script to match your file location).
  2. Create the output directory so OpenCV has somewhere to write: 
    mkdir -p data/videoframe
  3. Run the script from the helpers/ directory: 
    cd helpers
python convert_vid2image.py
  4. Frame images are written to data/videoframe/ and named sequentially — frame0.jpg, frame1.jpg, and so on.

​
How it works

The script opens the video with cv2.VideoCapture and enters a loop that advances one second on every iteration. Seeking is done by calling: 
cap.set(cv2.CAP_PROP_POS_MSEC, (count * 1000))
cv2.CAP_PROP_POS_MSEC sets the playback position in milliseconds, so multiplying the integer count by 1000 moves the read head to exactly count seconds into the video. The resulting frame is then encoded and saved as a JPEG via cv2.imwrite. The loop terminates when cap.read() returns ret = False, which happens once the end of the video is reached.
Video files are intentionally not checked into the repository due to their large file sizes. Only this helper script is provided. You must supply your own .avi episode files before running the extraction.
The output directory data/videoframe/ must exist before running the script. OpenCV’s cv2.imwrite will silently skip writing any frame if the destination directory is missing — no error is raised, and you will end up with an empty dataset.

Build docs developers (and LLMs) love

Get started for freeTalk to us