Data pre-processing

[priyansh1221]

Hello sir,
I'm getting difficulties in data pre-processing as the dataset is in png format and we are working on video. Please help me load data, pre-processing it, and preview it as a video or gif.
Thanks

[kaledhoshme123]

Hi @priyansh1221

• I will give you help on how you should go about reading the videos.
• The following example shows the mechanism by which a single video is read and the results are stored in an array.
• You should extend the work of the script so that you can read all the videos included in the dataset, in addition to the mechanism that you can skip a specified number of frames in order to read all videos of the same length (with the same number of frames), and you can do this by reading more about VideoCapture
  Good luck

import cv2
video_file_path = ''
image_height = 60
image_width = 60
video_reader = cv2.VideoCapture(video_file_path)
success, frame = video_reader.read()
previous = frame.copy()
frames_queue = []
while video_reader.isOpened():
    ok, frame = video_reader.read()
    if not ok:
        break
    diff = cv2.absdiff(frame, previous)
    f = cv2.resize(diff, (image_height, image_width))
    f = cv2.cvtColor(f, cv2.COLOR_BGR2GRAY)
    previous = frame.copy()
    normalized_frame = f / 255
    frames_queue.append(normalized_frame)
video_reader.release()