Why

I need to extract decoded frames from videos in order to feed DNN pipeline.

This is the answer package.

Install

pip install image2pipe

Usage

from multiprocessing import Queue
import logging
import cv2

import image2pipe

logging.basicConfig()


q = Queue(maxsize=4)
decoder = image2pipe.images_from_url(q, "shuttle-flip.mp4", fps="30", scale=(1000, 556))
decoder.start()

 for pair in yield_from_queue(q):
        fn, img = pair
        gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

        cv2.imshow("gray", gray)
        cv2.waitKey()
        cv2.destroyAllWindows()