I am conducting a test with two factorial parameters (x, y) in an image processing program, which has produced a number of test images as pngs of the same dimensions. For example, in the mtcars dataset, they could represent one illustrative car image for each of the cyl/carb combinations.

I would like to import these images into R and plot them in a facet grid with the respective parameter values on the axes (e.g. cyl and carb).

What's the best way to a) import the images into a tibble/df, b) plotting them as per above?

(If necessary, I'd be happy to update the question with example code once I know what package to use).

CodePudding user response：

If you have the images as png, you don't actually need to convert them to data frames. You can use geom_image from ggimage, which only requires the file path. Suppose I have two categories - "animal" and "color", with an image representing the conjunction of each unique "color" and "animal".

I need only do:

df <- data.frame(animal = rep(c("cat", "dog"), each = 3),
                 color = rep(c("red", "green", "blue"), 2),
                 img = path.expand(
                   c("~/redcat.png", "~/greencat.png", "~/bluecat.png",
                     "~/reddog.png", "~/greendog.png", "~/bluedog.png")))

library(ggimage)

ggplot(df)   
  geom_image(aes(x = 1, y = 1, image = img), size = 1)  
  facet_grid(color~animal, scales = "free")