Skip to content

Latest commit

 

History

History
31 lines (16 loc) · 795 Bytes

README.md

File metadata and controls

31 lines (16 loc) · 795 Bytes

Video Scene Detection and Segmentation

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.

Requirements

  • Python 3.x
  • OpenCV (cv2)
  • Scikit-Image (skimage)

Installation

pip install opencv-python scikit-image

Function Arguments

--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.