Z projections: when to use the different types

[Graham Wright]

Dear All,

I'm familiar with how z-projections work and why we use them (to visualize
3D data in 2D), however i would like to know more about when or why to use
the different types of projection (by which i'm referring to Minimum,
Maximum, Average, Sum slices, Standard deviation and Median). My
understanding is that Min is good for brightfield images and Max or Average
are good for fluorescence images.

Maybe the answer is simply that you use the one that looks best or most
representative of you data, but I don't feel satisfied by this and would
like some clearer idea.

Any help is much appreciated,
Graham

[Elizabeth Crowell]

Hello Graham,

There are a couple of additional things to consider when choosing which method
to use to make a z-projection.

The average z-projections can be very useful for visualizing 4-D time series in
2D.  We regularly use average projections to get an image of the path followed
by motile fluorescent particles, for example.
Using average rather than maximum or minimum can also help eliminate background
noise in projections of z-stacks.
However, if you need to make intensity measurements from the z-projection, you
must use "sum slices" to avoid introducing artifacts in the pixel intensity
values.  So it is important to know how the z-projection is made and how this
changes the raw data to be able to draw conclusions from your projected images.
For max/min, only the brightest/weakest intensity pixel in the stack is kept in
the final projection.  The names of the other types are fairly
self-explanatory.

I hope some of this information helps.

Elizabeth



Selon Graham Wright <[hidden email]>:

> Dear All,
>
> I'm familiar with how z-projections work and why we use them (to visualize
> 3D data in 2D), however i would like to know more about when or why to use
> the different types of projection (by which i'm referring to Minimum,
> Maximum, Average, Sum slices, Standard deviation and Median). My
> understanding is that Min is good for brightfield images and Max or Average
> are good for fluorescence images.
>
> Maybe the answer is simply that you use the one that looks best or most
> representative of you data, but I don't feel satisfied by this and would
> like some clearer idea.
>
> Any help is much appreciated,
> Graham
>

Elizabeth Crowell
Laboratoire de Biologie Cellulaire
INRA / Institut Jean-Pierre Bourgin
Route de Saint-Cyr
78026 Versailles cedex
FRANCE

Tel: [phone]
Fax: [phone]