What's New in ImageJ2?

Reproducible science with open tools

https://imagej.github.io/presentations/2017-06-01-imagej2-loci

Laboratory for Optical and Computational Instrumentation

LOCI develops advanced optical and computational techniques for imaging and experimentally manipulating living specimens.

Scopes

Science

Software

https://loci.wisc.edu/

What is ImageJ?

Fiji?

ImageJ2?

ImageJ1?

ImgLib2?

SciJava?

SCIFIO?

WAT

An application

A tool for image science

A reusable library


// Create a new ImageJ gateway with all available services.
ij = new ImageJ()

// Load a dataset from somewhere.
dataset = ij.io().open(imageSource)

// Filter the image.
fft = ij.op().fft(dataset)
lowpass(fft, 10)
result = ij.op().ifft(fft)

// Display the result.
ij.ui().show(result)

https://imagej.github.io/tutorials

An extensible collection
of services & plugins

A shared framework for
image processing

"Write once, run anywhere" image processing

Yes, ImageJ is all of that!

But most importantly...

ImageJ is a community!

Schindelin, J., et al. "The ImageJ ecosystem: an open platform for biomedical image analysis." Molecular reproduction and development 82.7-8 (2015): 518-529.

A sizable community...

^ ImageJ update checks, week of 2017-Feb-09

https://imagej.net/Usage

Motivation

Why use ImageJ? Why improve ImageJ?
Why not use Imaris and MATLAB for everything?
Why join this community?

Frugal: ImageJ is free and cross-platform
Pragmatic: You can fix bugs, add features
Scientific: See and understand how it works
Ethical: Fully reproducible, better science

Better science?

Why do science in the first place?

Improve the human condition
Discover new knowledge
Stand on each other's shoulders

21st century science requirements

Quantitative research
Keep track of everything

Information provenance

Remember what you did, and how
Explain it to someone else

Reproducibility

Validate others' work

Is this a good segmentation? (source)

Using 'Otsu' I find --- not a little difference --- but a big one. A little difference caused by the binning, a big one between implementations. Which is the 'real' Otsu threshold?
—Robert Atwood (source)

The Otsu algorithm had indeed an issue only with certain 16bit histograms where the computation of the variance could cause an overflow that carried on without triggering an error. I checked Emre Celebi's library (Fourier 0.8) and his C++ version does not have the problem so I replaced the previous version with Emre's. 8 bit images were not affected. I just checked and it finds the ball in your sample image now.

The Huang method is a different issue. The original code by Emre was rewritten in 2011 by Johannes Schindelin to use more efficiently the 16bit data and has a massive execution speed advantage. However I just noted that in some cases it does return exactly the same values as the original even in some 8bit images. I therefore added back the original code and kept Johannes' method as "Huang2" for the time being until I have time to look at it in more depth.

I also found that the Mean method (again in 16bits only) had a potential for overflow that could go without triggering errors, so I fixed that too.

—Gabriel Landini (source)

Sharing information

The Internet makes sharing easy

Wikipedia: public encyclopedia
Stack Exchange: public Q & A
GitHub: public source control
Twitter: public social interaction
Google: public etc.

"When in doubt, make it public."
—Jeff Atwood (co-creator of Stack Overflow)

Beyond open results

Publish a compendium, not just a result

Protocols & methodology
Raw data
Computer code

"An article about computational science in a scientific publication is not the scholarship itself, it is merely advertising of the scholarship."
—David Donoho, "Wavelab and Reproducible Research," 1995

Science Code Manifesto

Deep reproducibility demands the source code

"Software is a cornerstone of science. Without software, twenty-first century science would be impossible. Without better software, science cannot progress."
—Science Code Manifesto

And it is good for your career:

"Papers describing software published as open source are amongst the most widely cited publications (e.g., BLAST, and Clustal-W), suggesting many scientific studies may not have been possible without some kind of open software to collect observations, analyze data, or present results."
—Andreas Prlić & James Procter
"Ten Simple Rules for the Open Development of Scientific Software"

And it is good for your career:

"Science is hard enough already."
—Andreas Prlić & James Procter
"Ten Simple Rules for the Open Development of Scientific Software"

And it is good for your career:

Seriously.

Open software is a process

Ongoing open development
Improve software as a community
Open access resources
Responsive, reliable maintainers
Powerful collaboration tools (GitHub!)

^ Look familiar? ;-)

ImageJ Software Stack

Modular and extensible components

ImageJ2

=

SciJava
(base)

ImgLib2
(data model)

SCIFIO
(image I/O)

Ops
(image processing)

ImageJ Legacy
(compatibility)

https://imagej.net/Architecture

SciJava Pledge

A pledge to cooperate and reuse code

SciJava Component Collection

A collection of software components curated together

A promise of compatibility

Fiji Is Just ImageJ

A distribution of ImageJ for the life sciences
A community of ImageJ developers
Built on the ImageJ2 platform
Includes over 700 additional commands

https://fiji.sc/