New WSI viewer

[qigongFDA]

We might need a public WSI viewer for eeDAP. Then the users can enjoy the features in it (e.g. annotation) and we don't need to worry about color problem.

[qigongFDA]

I subscribed to openslide users email list and asked help from them:

My name is Qi Gong. I am working at the FDA. We have a project: evaluation environment for Digital and Analog Pathology (eeDAP). It is a system built by Matlab to help pathologist read the same field of view on WSI and microscope by registration between WSI and microscope stage. We published several papers about this system. (e.g. Registration accuracy between whole slide images and glass slides in eeDAP workflow)

Matlab has some limitations. Thus, we want to recreate eeDAP based on a public WSI viewer . We think openslide is a good candidate. We want to add a plugin to the viewer. It has several main features: 1. Get mouse clicked WSI x-y position. 2. Show certain FOV by command line. 3. Extract certain magnification level FOV and save to disk.

Does openslide have the plugin option? What language we should use to build the plugin (python, java, C...). Could anyone give us a guide about how to use openslide to view WSI, and how to create a basic plugin in it. Thank you.

[qigongFDA]

Reply from David de Mena:

Reply 1:
There is a python library base on openslide to create deepzoom images: https://github.com/openslide/openslide-python/blob/master/openslide/deepzoom.py Together with a openseadragon (or Googlemaps) JavaScript library you can navigate in the image with a level,x,y.

In an European project called AIDPATH Public Health System of Andalusia have created a DICOM WSI converter/viewer based on this system.

Reply 2:
For a WSI viewer, the best solution is based in a Web page: so you need a web server that serves images (like JIP2000) and a way to build “the pyramid” from original image. You can do it previously or on fly. Here http://libvips.blogspot.com.es/2013/03/making-deepzoom-zoomify-and-google-maps.html you have a good starting point.

Best options for viewers are openseadragon (https://openseadragon.github.io/docs/) and googlemaps,(https://developers.google.com/maps/documentation/javascript/?hl=es-419) both JS.

We have developed a DICOM converter/viewer research platform that we are preparing in an online testing. If you want, we can give you access to it.

Reply 3:
You can build an program, but the most command WSI Viewer are Web. So you need to embedded JavaScrip code into a web page and have server from a Web server (apache, or others).

So

1. create the pyramid structure with deepzoom (vips or python): that will create some folders and many frames
2. Insert into a webpage JavaScript command to “navigate” into the image (openseadragon or googlemaps)
3. link javascript with the structure of the pyramid with a dzi or indicating the folder or server in a url (like JIP2000)

There is no “magic” exe to make that… (unless you buy proprietary software).

Happy to help you!! If you need a formal consulting or software development we can also talk.

[qigongFDA]

Reply from Luca Lianas:

I’m Luca Lianas and I work at the Healthcare Flows group of the CRS4, a research center in Italy.
We’ve been working for a few years now on systems based on Digital Pathology images and, in particular, we developed a tool that can be considered as a plugin for Openmicroscopy’s OMERO (https://www.openmicroscopy.org/omero/ https://www.openmicroscopy.org/omero/) that basically act as a DZI adapter for OME’s platform.
Our plugin, called ome_seadragon (https://www.openmicroscopy.org/omero/ https://www.openmicroscopy.org/omero/) uses both OMERO and OpenSlide and rendering engine for the images and is based on the deepzoom project cited by David de Mena.
The plugin has both the back-end engine to render images in DZI format and the front-end viewer, which is OpenSeadragon, packet as a Javascript library.
The viewer has also drawing utilities developed by using Paper.js which make possible to the user to interact with the image and add 2D graphics on top of that as Regions of Interest.
ome_seadragon is now part of a few projects that we are running over the years, one of them is a e-learning platform called CyTest and if you want to “take a look” to the platform and how the viewer has been integrated with that you can read the publication on Research Gate https://www.researchgate.net/publication/305889225_CyTest_-_An_Innovative_Open-source_Platform_for_Training_and_Testing_in_Cythopathology https://www.researchgate.net/publication/305889225_CyTest_-_An_Innovative_Open-source_Platform_for_Training_and_Testing_in_Cythopathology
I think that ome_seadragon could be considered as a good example on how to use OpenSlide to convert images in DZI format and how to integrate it with an online viewer with ROI marking capabilities.
I hope that my two cents on the discussion were useful.

the link for the ome_seadragon tool is https://github.com/crs4/ome_seadragon https://github.com/crs4/ome_seadragon

[qigongFDA]

Reply from Thomas Melin:

You may find the following links interesting.
Desktop applications:
ASAP: https://github.com/GeertLitjens/ASAP
QuPath: https://qupath.github.io/

Web-based:
Cytomine: https://cytomine.coop/
Pathomation: http://free.pathomation.com/

[qigongFDA]

Reply from John

libvips can read images efficiently via the openslide library:

https://jcupitt.github.io/libvips/

The libvips "openslideload" operator was actually written by the
openslide developers.

You can load openslide images in nip2, the libvips GUI. nip2 is a
programmable GUI, so you can add features like transformation between
slide XY and pixel XY, if you want. It's open source and runs on
Linux, macOS and WIndows.

https://github.com/jcupitt/nip2

"vipsdisp" is a small image viewer that uses libvips to display images:

https://github.com/jcupitt/vipsdisp

So again it'll display openslide images. It has the nice feature that
screen painting is fully threaded and asynchronous.

[qigongFDA]

Download several viewers:
1.QuPath; 2.ASAP; 3.PMA.start; 4.Sedeen Viewer