Read, write and plot TIFFs

Volker J. Schmid, Bioimaging group, Department of Statistics, LMU Munich

2022-05-27

library(bioimagetools)

Read and write 3D RGB tiff stacks

readTIF() and writeTIF() read and write tiff stacks.

3D and color stacks are automatically recognized by readTIF(), if the meta information is correct. Available meta information is saved as attributes. readTIF() can also handle URLs. The function returns an array with range 0 to 1.

cell <- bioimagetools::readTIF("http://ex.volkerschmid.de/cell.tif")
## Warning in tiff::readTIFF(file, all = TRUE, info = TRUE, as.is = as.is, : TIFFReadDirectory: Unknown
## field with tag 50838 (0xc696) encountered
## Warning in tiff::readTIFF(file, all = TRUE, info = TRUE, as.is = as.is, : TIFFReadDirectory: Unknown
## field with tag 50839 (0xc697) encountered
print(dim(cell))
## [1] 512 512   3  52
print(attributes(cell)$dim)
## [1] 512 512   3  52
print(attributes(cell)$bits.per.sample)
## [1] 16
par(pty="s")
bioimagetools::img(cell, z=25 ,col="rgb")

writeTIF() writes out tiff stacks. Meta information in attributes are saved. The bits per sample can be controlled by an option.

writeTIF(cell, file="my_cell.tif")
## [1] 52

readClassTIF() is a special version of readTIF() for images with a very low number of intensity values, e.g. masks or segmented images.

bioimagetools::writeTIF(2*EBImage::thresh(cell[,,1,])+EBImage::thresh(cell[,,2,]), file="simple.tif")
## [1] 52
mysimple <- readClassTIF("simple.tif", n=3)
par(pty="s")
bioimagetools::img(mysimple[,,25],col="red",up=3)

## [1] TRUE
## [1] TRUE

Plotting

img() is a plotting function for images stored in an array.

par(pty="s")
bioimagetools::img(cell, z=25, col="rgb")

bioimagetools::img(cell, z=25, col="r")

bioimagetools::img(cell, z=25, col="grey")

Reading Bitmap files

Simple Bitmap files can be imported by readBMP().

bi<-readBMP(system.file(package = "bioimagetools","extdata", "V.bmp"))
par(pty="s")
image(bi, col=grey(seq(0,1,length=2^8)))

Single TIF slices

In order to work with single slices, you would like to select the slice from the array:

slice30<-cell[,,,30]

However, this will result in the wrong dimension. bioimagetools always needs 4d array:

dim(slice30) # only 3d
## [1] 512 512   3
dim(slice30) <- c(dim(slice30), 1)
dim(slice30) # 4d 
## [1] 512 512   3   1
img(slice30, z=1, col="rgb")

writeTIF(slice10,paste0("slice30.tif"))
## [1] 1

To read single RGB slices, you can use readTIFF. Make sure to add the option channels=3 for an RGB slice.

slice30<-readTIF("slice30.tif", channels=3)
dim(slice30)
## [1] 272 270   3   1

To stack single slices use abind from the abind package:

for (i in 1:52)writeTIF(cell[,,,i],paste0("slice",i,".tif")) # write slices
slice<-list()
for (i in 1:52)slice[[i]]<-readTIF(paste0("slice",i,".tif"), channels=3) # read slices
allslices<-slice[[1]] # slice 1
for (i in 2:52)allslices<-abind::abind(allslices, slice[[i]]) # stack slices 2 to 52
dim(allslices)
## [1] 512 512   3  52

The attributes will be off for the stacked array, so we need to add them from hand using one of the slices (not the dimension!):

attrib<-attributes(slice[[1]])
attrib<-attrib[names(attrib)!="dim"] # Remove dimension attribute 
attributes(allslices)<-c(attributes(allslices),attrib)