Look Up Tables

A quick check of the forum didn’t result in an answer, if this has been brought up, please forgive the duplication. 

Has it been suggested that there be a selection of Look Up Tables (LUTs) to be applied to the TEM black and white images?  The reason is to allow for colocalization of the cell membrane where the contrast is poor and the applied color by the player or as a starting point is covering the membrane.  A line of green+red would make a row of yellow pixels which would be much easier to see than the light grey where the membrane is not well defined.  Even with the transparency turned way down, on a 28" monitor, I have had some difficulty deciding if a membrane is obscured by the color or not. 

It works in confocal microscopy.

:slight_smile:

 

 

This is a bit over my head, so I’m going to pass it along to someone in the lab who’d be better able to answer!

Hi esotericman!


Thank you for the suggestion about the colors! I was testing a bit on the greyscale images to see if there is something that could help with blurry regions.

But maybe the most important answer for you first: You can hold the Shift key to completely hide your trace from the 2D view. That should allow you to see the membrane again :slight_smile:

Now, about the colors:

Lets just look at a single layer, without any modification. There are 141 different shades of grey in that layer, ranging from rgb(60,60,60) to rgb(214,214,214):

First I tried the green-yellow-red gradient. White becomes green (#00FF00), Black becomes Red (#FF0000). The colors can be easily calculated using the HSL color space and you get a really good looking transition from green over yellow to red. (Not sure if that is the best way, admittedly).
Result:

So… that didn’t work out well (or maybe I am just too used to the greyscale images) - I am looking at that one white-ish blob at the bottom, specifically. Not sure if the limitation is my monitor or my eyes. Maybe both. But I have a harder time to spot the differences.

Ok, next try: Stretching the existing values to the full range of 0-255. I scaled them even a bit more and then cut off the few outliers that won’t make a difference at all. By doing this we actually lose information due to rounding errors and of course the outliers that were cut. However, it should be easier to spot different shades now:

Yay, white blob!
Ok, let’s combine both:

With that you have the advantage of using the gradient (512 available colors instead of 256) and - thanks to the scaling - also make use of this full range. But it still looks a bit less clear

Phew, that was fun. There are still a lot of issues left and considerations to make, though. But I’d like to keep an eye on this topic. I find it really interesting. :slight_smile: