Stacks Image 351
Stacks Image 352
Stacks Image 353
Stacks Image 354
Stacks Image 355
Stacks Image 356
Stacks Image 357

Image Analysis

Such a hazard-filled subject, full of expectations and dizzying images - yet so full of pitfalls for the unwary.

If you're going to start analysing your images, it's likely you'll have to make a subjective call on at least some of the steps in the process. Typically they are 'where should I decide the boundary is?' and 'which intensity range should I choose to select my region of interest?'. These are important questions to answer well, as they have a strong influence on the final result.

Think of every image as a collection of numbers (they are after all 'digital' images) as then you will be more inclined to proceed with caution. If you see them merely as images, it's all too easy to become entirely subjective on whether they are acceptable or not.

The programs listed below are used at the OCCM and are a good starting point for many projects. Read through the descriptions to get an idea of what can be done. If you are unsure of what to do, or need help in developing an analytical method, please feel free to contact me at andrew.mcnaughton@otago.ac.nz.

For help on some of the methods relevant to work done at the OCCM, go to the 'How to …' page where you may find the procedure you want to do has already been documented.

Why so much open source software?

  • you can be trained and then use the same software on your own computer
  • online support forums are often more comprehensive and supportive than commercial software
  • often responsive to user requests for new or improved features
  • file formats are not proprietary and therefore unlikely to become unreadable if the company collapses
  • updates are free
  • the program is free!

Why not use programs like Photoshop?

Because they are often designed to make fabulous looking images - which they do very well. They are not designed to allow cold analytical objectivity to analyse your images. Avoid such programs except for a few minor house-keeping tasks such as cropping or tidying up rough edges. Don't say you haven't been warned.
Stacks Image 59

Fiji / ImageJ

A very underrated program, it can be expanded by installing any of the vast array of plugins and macros, which like the program itself are offered freely, courtesy of the open source software principle.

It's well worth checking the ImageJ
plugins page as it is full of nerdy extras developed by people just like you (face it, if you're reading this you're probably a nerd too). This means they have often had the same problems to solve and may well have done the work for you. Don't re-invent the wheel before checking to see if someone else hasn't already got the mags fitted.

Details of Image J can be found at Image J.

An alternative, and functionally identical program is Fiji, which comes ready packaged with many plugins and will search for updates and install them for you. It can be found under Fiji Project at the ImageJ links page - along with a lot of other useful programs.


The type of things typically done with Fiji / ImageJ include:

  • length, area, volume measurement
  • intensity and density levels
  • particle analysis
  • path tracing
  • volume rendering
Don't be put off by ImageJ's unassuming interface, it will perform many complex tasks and has the fabulous characteristic of being completely free and platform independent.
Stacks Image 77

3D Slicer

Designed for the medical profession, but perfectly happy dealing with any other type of 3D data such as confocal or µCT.

Be prepared for a fairly steep learning curve, but this really is a case of 'no pain, no gain'. Computers only make things easier if you want to stand still. If you want to progress, it takes the same amount of effort it always has.

3D Slicer runs on any operating system and is free - but please acknowledge the software if you use it.

Download 3D Slicer from their
website

More documentation is available from
here.
(You're going to need it.)
Stacks Image 269

Osirix

Designed for the medical profession, which means it's very good for viewing DICOM files and cataloging associated patient details in its database. That said, it's also very useful for viewing many other forms of 3D data sets, whether from the µCT or confocal.

Free for research purposes, Osirix does only run on Macintosh OS X, so if you are using a Windows OS then this little pearl of software isn't for you. It can be downloaded from Osirix's website in either 32 or 64 bit versions. The latter will cost you a few tens of dollars, but allows full use of your computer's RAM, which for large data sets can be essential.

Download Osirix from their
website

For some Windows alternatives, you could try looking at
http://alternativeto.net/software/osirix/ some may do what you need.

Stacks Image 67

Blender

Another open program which wildly exceeds what you'd expect for free software. Used primarily for 3D rendering and animations, Blender has a large and well informed community for support and inspiration.

Runs on Mac OS X, Linux and that other operating system, download it at Blender's website.

Don't expect to learn Blender in a hurry, the learning curve is close to vertical when you first begin. If you're thinking of using it, plan well ahead to allow time to become familiar with it. The reason it's difficult is it's capable of very complex tasks, not that the interface is poor, far from it.

Be very wary of programs such as this, they can be extremely useful, but remember they are designed for creating spectacular or fanciful images. This is not the same as presenting your data with objectivity and precision. Use it - but you've been warned - again.
Stacks Image 74

Amira

Not open source or even close to cheap, but it can be very good for volume and surface rendering - plus we have a couple of licences. It does have to be used on designated systems, so don't expect to do your image processing in the comfort of your own rat-infested student flat.