Research code on methods for segmentation of foreground image from green screen background.
Segmentation involves extracting specific parts of the image to make the image simpler or to extract a region of interest like a foreground object from the image.
This repsitory consists of general code that was used for foreground and background segmentation for the specific use case of images taken in a greenscreen under random lighting conditions. The general name of this technique is called as chroma-key segmentation.
The general methods of segmentation can be broadly classified into :
-
Classical computer-vision methods : Graphcut techniques , using color based segmentations
-
Deep Learning methods : U-Net for image segmentation etc.