I have an image where some pixels are missing. The missing pixels are represented as transparent. How do I interpolate transparent pixels based on their neighbours? In other words, how do I fill in values for them, based on the surrounding pixels?
Up till now, I have been using a hack to "smear" all the extant pixels to the right, filling in the missing ones overtop the transparent pixels, but this is crude. I have been using the Pillow and Proj libraries in Python and I see bilinear available when resizing but not for an existing image with missing pixels. The attached image on the left shows the transparent canvas with some pixels filled in, and others left transparent. On the right is the same image with a black BG, for clarity.