pavloshargan/VideoRotationAnalysis

This module estimates how much the video is rotated using RotNet CNN, which was pre-trained on GoogleStreetView data

VideoRotation estimation using mobilenet v2 custom classificator

Setup:

python3 -m pip install .

Pre-requirements:

ffmpeg is required

Usage:

import argparse
from videorotation import VideoRotationAnalysis

def main(video_path):
    with VideoRotationAnalysis(frames_per_video=12, frames_threshold=8) as analysis: #how many frames to check, and how many frames should be flipped to say that video is flipped
        result = analysis.check_if_upsidedown_for_video(video_path)
        print(result)

if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='Analyze video rotation.')
    parser.add_argument('video_path', type=str, help='Path to the video file.')

    args = parser.parse_args()
    main(args.video_path)