This script is designed to detect different scenes in a video and segment it into different clips. The detection is performed using Structural Similarity (SSIM) between frames. If the SSIM value falls below a specified threshold, a new scene is identified.
- Python 3.x
- OpenCV (
cv2) - Scikit-Image (
skimage)
pip install opencv-python scikit-image
--video_path: Path to the input video file.
--read_size: Number of frames to read in each batch (default: 500).
--min_interval: Minimum interval between scene borders (default: 200).
--ssim_threshold: SSIM threshold for scene change detection (default: 0.50).
--output_path: Directory to save the segmented clips.