OpenSlide.jl

This package is a Julia language wrapper for OpenSlide.

OpenSlide is a vendor-neutral library providing a common API to read virtual slide images generated by several different microscope vendors.

Julia is a new language for technical computing with familiar and script-friendly syntax, an extensive mathematical library and array-oriented computing built-in, and a high performance just-in-time compiler (< 2x of C code).

Installation

The package may be added by running the following command at the Julia prompt:

    julia> Pkg.add("OpenSlide")

Source code for the package is available from: http://github.com/ihnorton/OpenSlide.jl.git

The example image used in this notebook is available from: http://openslide.cs.cmu.edu/download/openslide-testdata/

High-level usage



In [1]:

    
# load the package
using OpenSlide

# construct path to test file inside of directory
# note: test file is not part of code distribution and must be downloaded separately.

imgfile = Pkg.dir("OpenSlide", "test", "CMU-1-Small-Region.svs")









    Out[1]:





"/home/isaiah/.julia/OpenSlide/test/CMU-1-Small-Region.svs"



In [2]:

    
# load the slide

@time img = OpenSlide.open_slide(imgfile)









    



elapsed time: 0.046075237 seconds (1129724 bytes allocated)






    Out[2]:





OpenSlideImage("/home/isaiah/.julia/OpenSlide/test/CMU-1-Small-Region.svs",1,[[2220,2967]],Ptr{openslide_t} @0x0000000005a6f000)



In [2]:

    
# read the entire level

@time imgdata = OpenSlide.read_slide(img);









    



elapsed time: 0.287445549 seconds (68811292 bytes allocated)



In [3]:

    
arraysize(imgdata)









    Out[3]:





(2220,2967,3)



In [13]:

    
using PyPlot

# Plot a sub-region
# We keep the figure small in order to reduce the size of the notebook, but larger figure size will work fine.
figure(figsize=[2,2])
imshow(imgdata[1200:1500,800:1100,:])









    
















    Out[13]:





PyObject <matplotlib.image.AxesImage object at 0xdb76310>



In [6]:

    
# read a given extent and level (shown for example, but note that this image has only one level!)

@time imgdata2 = OpenSlide.read_slide(img; origin = [1200,800], extent = [300,300], level = 1);









    



elapsed time: 0.001220545 seconds (901960 bytes allocated)



In [12]:

    
figure(figsize=[2,2])
imshow(imgdata2)









    
















    Out[12]:





PyObject <matplotlib.image.AxesImage object at 0xa9ce510>



In [14]:

    
# get image properties

props = OpenSlide.properties(img)

props["tiff.ResolutionUnit"]









    Out[14]:





"inch"

Low-level API

The package also includes wrappers for all of the non-deprecated functions in openslide.h (see src/libopenslide.jl). All functions have been stripped of the openslide_ prefix, except for openslide_open (so as not to conflict with Julia's open function).

Function listing

    get_property_names,
    get_property_value,
    can_open,
    openslide_open,
    get_level_count,
    close_slide,
    get_level0_dimensions,
    get_level_dimensions,
    get_level_downsample,
    get_best_level_for_downsample,
    read_region,
    get_associated_image_names,
    read_associated_image,
    get_error,
    openslide_version

Notes

Julia's default array ordering is column-major. Data from OpenSlide must be appropriately converted, see libopenslide.jl::_rgbfromrowflat