Skip to content

Homemade green screen processor putting Shia LaBeouf in any photo.

License

Notifications You must be signed in to change notification settings

xcollantes/green-screen-chroma-key-shia

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Green Screen Processor for Shia LaBeouf

Take this image of Shia LaBeouf with a green screen behind him from the famous meme where Shia stands in front of a green screen giving a strangely effective motovational speech.

If you haven't seen it, check it out. You'll be glad you did.

Shia LaBeouf

Superimpose Shia on the below image of me and friends.

For Good

How It Works

  1. Import pillow library or Python Image Library must be used to manipulate images on a pixel level.

  2. Each pixel of the image containing the green screen is parsed and compares RGB value with a threshold in order to determine the 'greenness' of the pixel.

  3. Once the 'greenness' of the pixel is determined to be a green screen element, replace the pixel with a corresponding pixel from the background image.

  4. After this is repeated for all pixels (Shia LaBeouf image is 509x905, which means 460,645 pixels), save the image as a standalone.

  5. The threshold is the heart of the algorithm, therefore outputing the RGB values for each pixel from the Shia LaBeouf image into a CSV could provide insight to an improved algorithm.

  6. Voilá, we have our final product. I don't seem to have made the final cut but that's okay.

Shia LaBeouf

"Don't let your dreams be dreams!" Shia LaBeouf

Next Steps

Notice a green outline on our hero, Shia LaBeouf, in the final image.
Fine tuning the chroma key algorithm to detect a greater range of greens could result in a cleaner image.

Libraries Used

os for writing our RGB values to find threshold in foreground image.
pillow or Python Image Library to manipulate images on a pixel level.

Releases

No releases published

Packages

No packages published

Languages