Breakthrough/PySceneDetect

:movie_camera: Python and OpenCV-based scene cut/transition detection program & library.

What it solves

PySceneDetect is a tool for automatically detecting scene cuts in videos. It eliminates the manual effort of identifying where one shot ends and another begins, allowing users to split videos into individual scenes or save representative frames from each cut.

How it works

The tool analyzes video content to identify transitions. It offers three primary detection methods:

  • ContentDetector: Performs content-aware scene detection to find fast cuts.
  • AdaptiveDetector: A two-pass detector designed to handle fast camera movement more effectively.
  • ThresholdDetector: Specifically handles fade-in and fade-out events.

Users can interact with the tool via a command-line interface (CLI), a Docker image, or a Python API for integration into larger pipelines. It can use ffmpeg or mkvmerge to physically split the video files based on the detected scenes.

Who it’s for

This tool is for video editors, developers building video processing pipelines, and researchers who need to programmatically analyze and segment video content.

Highlights

  • Multiple Detection Algorithms: Includes specialized detectors for fast cuts, camera movement, and fades.
  • Funding-Free Integration: Works as a standalone CLI tool or as a Python library.
  • Automated Splitting: Integrates with ffmpeg and mkvmerge to automatically cut videos into separate files.
  • Flexible Output: Ability to save specific frames from each detected scene for quick review.

Related

  • Project
    ssrajadh/sentrysearch

    A semantic video search tool that lets you find specific events in footage using natural language or images and automatically trims the matching clips.

  • Project
    Breakthrough/DVR-Scan

    A command-line and GUI application that automatically detects motion events in video files and saves each event as a separate clip.

  • Project
    HUANGCHIHHUNGLeo/claude-real-video

    A local video processing pipeline that extracts meaningful keyframes and transcripts for LLMs, replacing fixed-interval sampling with scene-change detection and deduplication.

  • Project
    leblancfg/autocrop

    A Python-based tool that uses the YuNet neural network to automatically crop images around the largest detected face, ideal for profile pictures and ID cards.

  • Project
    0x09/resdet

    A command-line tool and C library that detects upscaling in images and videos to identify their original source resolution using frequency domain analysis.