I'm trying to improve the identification of a facial recognition in a photo, because sometimes no face is detected using Microsoft's Face API. In this scenario I want to try to improve the quality of the photo. When I do this manually in Microsoft's Windows 10 Photo app by sliding the Clarity bar, I can now detect a face.
Using the C# library called ImageProcessor, I figured Microsoft's clarity feature is improving contrast and possibly GaussianSharpen. I have used a wide variety of combinations of these two in ImageProcessor, but I don't get the same results.
To view someone else who ran into a similar problem, you can click here: Windows 10 Photos App "Clarity" option algorithm
However, I do not have the reputation to ask in a comment if this had been figured out.
imageFactory.Load(img)
.Contrast(100)
.GaussianSharpen(new
ImageProcessor.Imaging.GaussianLayer(10, 1.5, 10))
.Save(sharpenedImage);
Due to not wanting to post pictures of people's faces online, I cannot provide a whole lot of detail, but would appreciate any insights or direction. The link above does have examples of what the clarity slider does.
In my manual testing I slide clarity to 100.
My C# code fixes some images but not all. Microsoft's clarity feature fixes all.
