hecate  by yahoo

Video processing CLI tool for thumbnails, GIFs, and summaries

created 9 years ago
493 stars

Top 63.5% on sourcepulse

GitHubView on GitHub
Project Summary

HECATE is a C++ video processing library designed to automatically generate thumbnails, animated GIFs, and video summaries. It targets developers and researchers working with video content who need to extract visually appealing and relevant clips or images without manual intervention. The library leverages computer vision techniques to identify high-quality frames and segments, simplifying the creation of engaging video previews and summaries.

How It Works

HECATE processes videos by first detecting shot boundaries and extracting keyframes. It then applies aesthetic and relevance criteria, detailed in a cited research paper, to select the most suitable frames for thumbnails and segments for GIFs and summaries. This approach aims to produce outputs that are both representative of the video's content and visually pleasing, automating a traditionally labor-intensive process.

Quick Start & Requirements

  • Installation: Clone the repository, configure Makefile.config with OpenCV include/library paths, and run make all followed by make distribute.
  • Prerequisites: OpenCV library with FFMPEG support.
  • Usage: Executable hecate with options for input video (-i), output directory (-o), frame subsampling (-s), number of thumbnails (-njpg), number of GIFs (-ngif), summary length (-lmov), and output dimensions.
  • Example: ./distribute/bin/hecate -i examples/video.mp4 --generate_jpg --njpg 3
  • Documentation: GitHub Repository

Highlighted Details

  • Automatic generation of thumbnails, animated GIFs, and video summaries.
  • Includes shot boundary detection and keyframe extraction.
  • Selects frames based on aesthetic and relevance criteria.
  • Customizable output parameters (number of items, dimensions, summary length).

Maintenance & Community

  • Developed and maintained by Yahoo Research.
  • Primary developer: Yale Song.
  • Cites research paper: "To Click or Not To Click: Automatic Selection of Beautiful Thumbnails from Videos" (CIKM 2016).

Licensing & Compatibility

  • Licensed under the Apache 2.0 License.
  • Permissive license suitable for commercial use and integration into closed-source projects.

Limitations & Caveats

The library was last updated in 2016, indicating potential compatibility issues with modern systems or dependencies. The build process relies on manual configuration of Makefiles and external library paths, which can be complex.

Health Check
Last commit

2 years ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
0
Star History
4 stars in the last 90 days

Explore Similar Projects

Feedback? Help us improve.