Image Processing Operations inside R shiny server.R - r

Question

I'm currently working on image processing application using R Shiny It uploads an image using file upload and then I need to read the image to do the image processing operations. server.R file is as follows.
library(shiny)
library(EBImage)
library(imager)
library(jpeg)
function(input, output) {
observe({
file_path <- input$files
if (is.null(file_path))
return(NULL)
file_path$datapath <- gsub("\\\\", "/", file_path$datapath)
img <- readImage(file_path$datapath)
equalized <- equalize(img,range = c(0, 1), levels = 256)
output$text <- renderText({
file_path$datapath
})
output$img <- renderImage({
list(src = file_path$datapath,
contentType = "image/jpg",
width = "50%",
height = "auto",
alt = "This is alternate text")
})
})
}
But this gives me the following error.
Warning: Error in readImage: Please supply at least one filename.
Stack trace (innermost first):
57: readImage
56: observerFunc

Answer 1

I managed to plot an equalized image using raster method. Here are some tips/tricks:
You put everything inside an observer which is a pretty bad idea, so I got rid of that.
Use req() when checking whether a file is uploaded, UI is rendered, etc instead of an if statement. `
if (is.null(file_path)) return(NULL)
There is no need to assign input$files to a variable, you can call input$files$datapath. Also gsub() is not needed in this case.
file_path <- input$files
file_path$datapath <- gsub("\\\\", "/", file_path$datapath)
equalized is calculated, but you don't use it anywhere.
Solution
Checking with req() whether a file is uploaded.
Get the extension of the file (splitting by ., getting the last element)
Plot the equalized image using the display() function with method = "raster".
Print datapath which points to a temp dir/file
See:
library(shiny)
library(EBImage)
library(imager)
library(jpeg)
ui <- fluidPage(
fileInput("files", "Upload a file"),
plotOutput("img"),
textOutput("txt")
)
server <- function(input, output) {
output$img <- renderPlot({
req(input$files)
st <- strsplit(input$files$name, split = "[.]")[[1]]
extension <- st[length(st)]
display(equalize(readImage(input$files$datapath, type = extension), range = c(0, 1), levels = 256), method = "raster")
})
output$txt <- renderText({
input$files$datapath
})
}
shinyApp(ui, server)

Answer 2

Using EBImage we can load an image into Rshiny and use it for further processing. Below code allows the user to upload an image and then same is displayed on shiny screen back.
library(shiny)
library(EBImage)
upload_image <- list()
ui <- fluidPage(
fileInput("file1", "Upload an Image"),
plotOutput("img")
)
server <- function(input, output) {
output$img <- renderPlot({
req(input$file1)
upload_image[[1]] <- readImage(input$file1$datapath)
plot(upload_image[[1]])
})
}
shinyApp(ui , server)