Skip to content

Latest commit

 

History

History
32 lines (25 loc) · 1.54 KB

README.md

File metadata and controls

32 lines (25 loc) · 1.54 KB
Raw

Project title

Chroma Keying, an implemention of chroma keying using purely image processing techniques.

Motivation

Chroma keying is a popular technique used in photography/videography to composite multiple images and video stream together. Chroma keying feature is offered in a lot of video post-processing products such as Adobe After Effects, Premiere Pro, Apple Final Cut Pro X,...However, these products are expensive, therefore, this project could be a free alternative for students doing school works or learning.

Tech/framework used

  • Python
  • OpenCV
  • Image processing

Features

  • Allow either automatic or manual background detection/removal
  • User-friendly manual background detection by using scribbles and trackbars
  • Accurate background removal in automatic mode

How to use?

TODO

Limitations

  • Required the background to be uniform in color, smooth in texture.
  • Automatic background detection only works when the background covers the majority of the area in the image.
  • Works poorly with partially transparent foreground object since the alpha channel estimation functionality is missing.

What's next?

  • Implement alpha channel estimation for better background removal consistency
  • Implement seamless cloning of the foreground into different background image/video
  • Take the idea from this project and create a

Credits

Big thanks to @spmallick for providing resourceful learning materials. This project is an implementation of this research paper: https://pdfs.semanticscholar.org/bf62/cae2be4b4785f4a54ba9d9ef54e0ba61068c.pdf