Skip to content

Latest commit

 

History

History
89 lines (67 loc) · 3.58 KB

README.md

File metadata and controls

89 lines (67 loc) · 3.58 KB
Raw

Inpainting

license Build Status Drone Status Donate Tweet

.NET implementation of content-aware fill (also known as inpainting or image completion) in image processing domain.

What is it for?

Content aware fill is used to fill in unwanted or missing areas of photographs. See an example of such fill below where we don't want to have the man in blue t-shirt on our picture:

Original image Processed image
t009 r009

Try in Docker

docker run -p 5000:80 -it --rm zavolokas/inpainter:latest

Navigate to http://localhost:5000

How to use it?

var inpainter = new Inpainter();
var result = inpainter.Inpaint(imageArgb, markupArgb, donors);
result
    .FromArgbToBitmap()
    .SaveTo(resultPath, ImageFormat.Png)
    .ShowFile();

The Inpainter takes as input

  • an image to inpaint
  • a simitransparent image with a mask
  • optionally it takes a set of simitransparent images that define donor areas for the parts of the area to inpaint.
  • optionally it takes an instance of settings.

Note: the images are not GDI+ images but images in an internal format and can be obtained from GDI+ Bitmaps using extensions.

Examples

Original Markup Process
t009 m009 p009
t020 m020 p020
t023 m023 p023
t058 m058 p058
t067 m067 p067

Settings

The execution of the algorithm can be customized by adjusting the settings.

  • MaxInpaintIterations: determines how many iterations will be run to find better values for the area to fill. The more iterations you run, the better result you'll get.
  • PatchDistanceCalculator: determines algorithm to use for calculating a metrics how much one color is different from another. Possible values are:
    • Cie76 - fastest
    • Cie2000 - more accurate

Donors

Credits

The implementation is based on following publications:

  • Yonatan Wexler, Eli Schechtman and Michal Irani Space-time completion of video IEEE. Trans. Pattern Analysis and Machine Intelligence, 29 (2007)
  • Connelly Barnes, Eli Shechtman, Adam Finkelstein, and Dan B Goldman. PatchMatch: A Randomized Correspondence Algorithm for Structural Image Editing. ACM Transactions on Graphics (Proc. SIGGRAPH) 28(3), August 2009