ImageJ

Top

00:00:15.00 Hi, my name is Kevin Eliceiri of the University of Wisconsin at Madison
00:00:18.01 and the ImageJ team.
00:00:19.22 And in this part of our series on image analysis,
00:00:23.14 we wanted to take time to profile various
00:00:26.08 open-source image analysis tools,
00:00:27.26 one of them being ImageJ.
00:00:29.19 And ImageJ has been a long-standing tool
00:00:32.24 that is available freely and open in the bioimaging community.
00:00:36.06 And it's been in development since 1997,
00:00:38.24 all the way back to its original roots at NIH,
00:00:41.23 where it was being developed and had a very strong...
00:00:44.23 thousands of users base,
00:00:47.09 where it could easily extend and develop this program.
00:00:49.27 And as the program became more popular,
00:00:51.16 and biologists with very little programming background
00:00:53.25 were able to extend this program,
00:00:55.27 my group saw the need to extend it with our collaborators,
00:00:58.09 where we developed ImageJ2,
00:00:59.28 which is basically the next generation of ImageJ,
00:01:01.13 to allow it to actually have
00:01:03.23 an easier ability to collaborate with other tools,
00:01:06.26 and interoperate,
00:01:08.14 more modern plugins,
00:01:09.29 but also compatible with old plugins.
00:01:11.22 And really allowing new features, such as better support for big data,
00:01:15.11 better updating,
00:01:17.03 and better scripting.
00:01:18.10 And then Fiji is ImageJ 1 and 2, with batteries included.
00:01:22.08 It's basically a way of getting all the goodness of ImageJ,
00:01:24.20 but for a microscopy audience.
00:01:27.06 And allowing you to actually download this
00:01:29.26 with active maintenance and updates released,
00:01:31.24 and we recommend Fiji as the preferred version...
00:01:34.13 version of the viewers of the series because it's really for microscopists.
00:01:37.17 So, even though it could be a little confusing,
00:01:40.23 thinking about ImageJ1, ImageJ2, and Fiji,
00:01:42.12 just really think ImageJ and Fiji,
00:01:44.13 because Fiji contains all the goodness of ImageJ
00:01:46.17 that you would want to use for image analysis of bioimaging.
00:01:50.11 So, what is ImageJ?
00:01:51.19 Because it actually has continued to evolve.
00:01:53.23 It is, foremost, an application.
00:01:57.09 It's a downloadable application,
00:01:59.11 very easy to use, very simple interface, as you can see here.
00:02:02.29 And it's designed to do image analysis.
00:02:05.21 But it's not designed to compete with
00:02:08.11 very powerful programs like Photoshop.
00:02:10.05 It's really a tool for doing science.
00:02:11.20 Photoshop is wonderful for making posters and doing image analysis,
00:02:15.03 particularly for commodity work.
00:02:17.12 But when we're thinking about science and analyzing images,
00:02:19.16 particularly 2D and 3D,
00:02:21.16 ImageJ is for that.
00:02:23.02 It's designed to be very transparent.
00:02:25.02 It's really designed for open science,
00:02:26.23 so you know what is happening to your data.
00:02:28.21 It is an application that's hopefully easy to use,
00:02:30.26 and it's really designed to be well-documented,
00:02:33.22 with a simple interface.
00:02:35.10 And it's also meant to be a tool for image science.
00:02:37.19 What we mean by that is
00:02:40.07 really having the tools that you need
00:02:42.19 for microscopy in particular.
00:02:44.03 How you can open up your data,
00:02:46.13 for example, using the powerful tool Bio-Formats,
00:02:48.11 that our lab and others helped develop,
00:02:50.05 where you can open up different formats from different microscopes.
00:02:52.21 Being able to do a very important, advanced analysis
00:02:56.29 that we have covered in the overall
00:02:59.23 basics of image analysis series,
00:03:01.05 such as registration, segmentation,
00:03:02.29 measurements, 3D volume rendering,
00:03:04.21 and many more.
00:03:06.06 And there are a number of features
00:03:08.20 and hundreds and hundreds of plugins
00:03:10.14 that have been contributed from around the world
00:03:11.24 to help you do your science.
00:03:13.05 But there are a few plugins that I wanted to highlight
00:03:14.24 just to give you a taste of the program.
00:03:16.22 One of them is Trainable Weka.
00:03:18.12 As my colleague Anne Carpenter talked about
00:03:20.15 in her part on machine learning,
00:03:21.26 it's very important to have
00:03:23.29 machine learning-based methods.
00:03:25.01 And one of them is the Weka segmentation plugin
00:03:27.20 that allows you to have a very user-friendly tool
00:03:30.20 to integrate machine learning segmentation into your analysis workflow,
00:03:34.17 as shown here.
00:03:35.24 And the ability to do Weka in ImageJ has become very popular
00:03:38.02 for doing automated segmentation.
00:03:41.00 Another plugin I want to highlight
00:03:43.04 is the widely used TrackMate.
00:03:44.17 This is a wonderful open-source plugin for ImageJ,
00:03:46.21 and also part of Fiji,
00:03:48.18 that's a great tool for tracking cells and objects.
00:03:51.19 As we covered in the tracking part of our video series
00:03:54.02 on image analysis basics,
00:03:56.02 tracking is a very powerful method
00:03:58.07 to track 2D and 3D objects
00:04:00.08 -- including viruses, particle, cells --
00:04:03.27 as they go over time and space.
00:04:05.10 TrackMate is a very well-documented and well-published program
00:04:09.23 to allow you to do this easily.
00:04:11.07 You can actually run existing algorithms
00:04:13.10 and get data fairly seamlessly out of the program.
00:04:15.21 But in the spirit of ImageJ,
00:04:17.05 you can also extend it.
00:04:18.08 And it's actually fairly easy to add your own tools to this,
00:04:20.16 and groups have added their own linking algorithms,
00:04:22.18 for example,
00:04:24.10 to do their own tracking.
00:04:25.15 So, I highly encourage you to look at TrackMate.
00:04:27.03 Another tool -- it's one of our newest additions --
00:04:29.19 is SciView.
00:04:31.13 One of the classic things that you want to do
00:04:33.17 in programs like ImageJ
00:04:35.02 is be able to visualize your event.
00:04:36.17 As I covered briefly in the visualization part of an earlier talk
00:04:39.11 I gave as part of this series,
00:04:41.03 we really want to understand a 3D object,
00:04:44.16 what happens in 3D,
00:04:46.03 and be able to visualize a 3D process beyond just slices of the stack.
00:04:48.11 We want to see it in its totality.
00:04:50.10 3D rendering allows us to do this,
00:04:52.18 and one of the newer things we want to do in ImageJ and Fiji
00:04:55.26 is actually have better views.
00:04:57.00 SciView allows you to have high performance with big data,
00:05:00.12 where you can actually look at the scene
00:05:03.13 as a rendering back-end,
00:05:05.06 and support new features including virtual reality,
00:05:07.28 where you can roam and visualize your data in 3D.
00:05:10.26 And SciView is also part of Fiji, now,
00:05:12.17 and a very powerful way to render and visualize your data.
00:05:15.22 I do also want to point out that ImageJ goes beyond
00:05:18.26 just being a simple user-based interactive software package.
00:05:21.12 It has built-in usability for users
00:05:24.22 to be able to modify and write scripts and macros.
00:05:27.12 But it's also designed to be for developers,
00:05:30.11 including collaborators -- such as Anne Carpenter of ImageJ, Jason Swedlow at OME,
00:05:34.17 other speakers in this series --
00:05:36.13 where they've been able to extend their programs
00:05:38.19 by using libraries from ImageJ, and vice versa.
00:05:40.16 We want to have ImageJ to be easy to reuse and share
00:05:44.07 beyond one user and one application.
00:05:48.04 And that means we actually have a shared framework.
00:05:50.11 We've actually made it easy for algorithms to be shared,
00:05:52.21 as you can see here,
00:05:54.02 with a number of different toolkits from other developers and labs,
00:05:56.25 being able to extend our functionality.
00:05:59.01 And one of them is Ops.
00:06:00.17 Ops is meant to be a way of sharing image processing operations
00:06:03.04 between different tools
00:06:07.14 so that we can actually enable programmers to develop their own image processing
00:06:10.03 in a different tool,
00:06:11.17 and share and reuse in this framework we call SciJava.
00:06:14.01 And this can include ImageJ, CellProfiler,
00:06:16.12 KNIME, OMERO, and others,
00:06:18.16 to share and reuse so that together
00:06:21.03 we can actually have a better workflow,
00:06:22.27 and also use our energy and activities in the most...
00:06:25.12 best way possible...
00:06:26.20 best way possible.
00:06:28.12 And so we've worked hard in ImageJ
00:06:30.14 to interoperate and have integrations
00:06:32.12 with many other tools, including CellProfiler, KNIME, OMERO,
00:06:36.01 libraries such as NumPy, ITK.
00:06:39.09 And really the idea is to create a unified resource --
00:06:43.09 we all work together.
00:06:45.04 And this is not only seen by our interactions in the code,
00:06:47.09 but also in the forum that we mentioned in another part of the video series,
00:06:50.15 where we actually can ask questions online
00:06:52.27 and work together, both users and developers.
00:06:56.08 Who uses ImageJ?
00:06:57.18 It's a pretty diverse base.
00:06:58.29 As you can see here
00:07:00.16 -- now, these are all equally important to us --
00:07:02.07 we always start with the very important novice users and power users;
00:07:07.04 but also scripters and hackers,
00:07:08.28 people that want to modify the code,
00:07:10.22 that maybe are in between biology and...
00:07:13.28 and hacking or scripting;
00:07:15.18 or professional developers that will actually want to add new functionality.
00:07:18.23 And that's really the goal we have,
00:07:20.12 is to span the gamut of different applications.
00:07:23.20 It's also important to always understand,
00:07:25.14 what are the values that drive the project?
00:07:27.00 What drives the ImageJ project
00:07:29.12 is that we want to be functional.
00:07:30.19 We want to be extensible.
00:07:31.19 We want to interoperate with others.
00:07:33.03 We want to be accessible.
00:07:34.10 We want to make sure you could download our program freely
00:07:36.11 and use it without cost,
00:07:38.17 but also know what's happening.
00:07:39.27 You want to know... transparency.
00:07:41.21 You want to be able to know that if you download this code
00:07:43.19 and your colleague downloads it,
00:07:44.23 you get the same result, that it's reproducible.
00:07:46.19 That really is an important part of good science.
00:07:49.03 Be compatible with as many different operating systems and different workflows,
00:07:51.27 whether you're running it on the cloud
00:07:53.21 or you're running it on your computer.
00:07:55.17 And really, we want to think about community.
00:07:57.15 So, again, that forum that we just started is so important.
00:08:00.05 We really want to make sure we're sharing
00:08:02.19 our problems, our successes, together,
00:08:05.16 so that we can progress on our different projects
00:08:07.13 in the most efficient manner.
00:08:09.00 To learn more, I encourage you to go to our website,
00:08:11.01 which is shown here.
00:08:12.16 But I also really encourage you to go to the forum.
00:08:15.00 That really is the best place
00:08:16.28 not only to see our tool but all the other tools that are out there
00:08:20.04 in this particular domain.
00:08:21.16 And we all are trying to work together,
00:08:23.08 so please ask your questions.
00:08:24.22 Don't suffer in silence.
00:08:26.08 If you have a problem, please ask on the forum.
00:08:28.02 Thank you for your time,
00:08:29.15 and I encourage you to watch the other videos in the series,
00:08:31.06 including the one on CellProfiler,
00:08:32.24 but also the basics,
00:08:35.07 that cover the fundamentals of image analysis in the other parts of the series.
00:08:37.24 Thank you.

This Talk

Speaker: Kevin Eliceiri

Audience:

Educators of Adv. Undergrad / Grad
Researcher
Educators

Recorded: April 2019

Talk Overview

Since 1987, different versions of ImageJ have been used by scientists to analyze biological images. In this talk, Dr. Kevin Eliceiri provides an overview of ImageJ, explains how ImageJ has evolved through time, and demonstrates major functionalities of this open-source software.

Concepts:
ImageJ, Fiji, ImageJ2, Weka Segmentation, TrackMate, SciView

Speaker Bio

Kevin Eliceiri

Dr. Kevin Eliceiri is an Associate Professor of Medical Physics and Biomedical Engineering and director of the Laboratory for Optical and Computational Instrumentation (LOCI) at the University of Wisconsin-Madison. Eliceiri completed his bachelor’s and doctoral degree in Biotechnology and Biomedical Engineering at the University of Wisconsin in Madison. In 2008, he founded LOCI and started… Continue Reading